fs/afs/cell.c

/* AFS cell and server record management
 *
 * Copyright (C) 2002 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 License
 * as published by the Free Software Foundation; either version
 * 2 of the License, or (at your option) any later version.
 */

#include <linux/module.h>
#include <linux/slab.h>
#include <rxrpc/peer.h>
#include <rxrpc/connection.h>
#include "volume.h"
#include "cell.h"
#include "server.h"
#include "transport.h"
#include "vlclient.h"
#include "kafstimod.h"
#include "super.h"
#include "internal.h"

DECLARE_RWSEM(afs_proc_cells_sem);
LIST_HEAD(afs_proc_cells);

static struct list_head afs_cells = LIST_HEAD_INIT(afs_cells);
static DEFINE_RWLOCK(afs_cells_lock);
static DECLARE_RWSEM(afs_cells_sem); /* add/remove serialisation */
static struct afs_cell *afs_cell_root;

#ifdef AFS_CACHING_SUPPORT
static cachefs_match_val_t afs_cell_cache_match(void *target,
						const void *entry);
static void afs_cell_cache_update(void *source, void *entry);

struct cachefs_index_def afs_cache_cell_index_def = {
	.name			= "cell_ix",
	.data_size		= sizeof(struct afs_cache_cell),
	.keys[0]		= { CACHEFS_INDEX_KEYS_ASCIIZ, 64 },
	.match			= afs_cell_cache_match,
	.update			= afs_cell_cache_update,
};
#endif

/*
 * create a cell record
 * - "name" is the name of the cell
 * - "vllist" is a colon separated list of IP addresses in "a.b.c.d" format
 */
int afs_cell_create(const char *name, char *vllist, struct afs_cell **_cell)
{
	struct afs_cell *cell;
	char *next;
	int ret;

	_enter("%s", name);

	BUG_ON(!name); /* TODO: want to look up "this cell" in the cache */

	/* allocate and initialise a cell record */
	cell = kmalloc(sizeof(struct afs_cell) + strlen(name) + 1, GFP_KERNEL);
	if (!cell) {
		_leave(" = -ENOMEM");
		return -ENOMEM;
	}

	down_write(&afs_cells_sem);

	memset(cell, 0, sizeof(struct afs_cell));
	atomic_set(&cell->usage, 0);

	INIT_LIST_HEAD(&cell->link);

	rwlock_init(&cell->sv_lock);
	INIT_LIST_HEAD(&cell->sv_list);
	INIT_LIST_HEAD(&cell->sv_graveyard);
	spin_lock_init(&cell->sv_gylock);

	init_rwsem(&cell->vl_sem);
	INIT_LIST_HEAD(&cell->vl_list);
	INIT_LIST_HEAD(&cell->vl_graveyard);
	spin_lock_init(&cell->vl_gylock);

	strcpy(cell->name,name);

	/* fill in the VL server list from the rest of the string */
	ret = -EINVAL;
	do {
		unsigned a, b, c, d;

		next = strchr(vllist, ':');
		if (next)
			*next++ = 0;

		if (sscanf(vllist, "%u.%u.%u.%u", &a, &b, &c, &d) != 4)
			goto badaddr;

		if (a > 255 || b > 255 || c > 255 || d > 255)
			goto badaddr;

		cell->vl_addrs[cell->vl_naddrs++].s_addr =
			htonl((a << 24) | (b << 16) | (c << 8) | d);

		if (cell->vl_naddrs >= AFS_CELL_MAX_ADDRS)
			break;

	} while(vllist = next, vllist);

	/* add a proc dir for this cell */
	ret = afs_proc_cell_setup(cell);
	if (ret < 0)
		goto error;

#ifdef AFS_CACHING_SUPPORT
	/* put it up for caching */
	cachefs_acquire_cookie(afs_cache_netfs.primary_index,
			       &afs_vlocation_cache_index_def,
			       cell,
			       &cell->cache);
#endif

	/* add to the cell lists */
	write_lock(&afs_cells_lock);
	list_add_tail(&cell->link, &afs_cells);
	write_unlock(&afs_cells_lock);

	down_write(&afs_proc_cells_sem);
	list_add_tail(&cell->proc_link, &afs_proc_cells);
	up_write(&afs_proc_cells_sem);

	*_cell = cell;
	up_write(&afs_cells_sem);

	_leave(" = 0 (%p)", cell);
	return 0;

badaddr:
	printk(KERN_ERR "kAFS: bad VL server IP address: '%s'\n", vllist);
error:
	up_write(&afs_cells_sem);
	kfree(cell);
	_leave(" = %d", ret);
	return ret;
}

/*
 * initialise the cell database from module parameters
 */
int afs_cell_init(char *rootcell)
{
	struct afs_cell *old_root, *new_root;
	char *cp;
	int ret;

	_enter("");

	if (!rootcell) {
		/* module is loaded with no parameters, or built statically.
		 * - in the future we might initialize cell DB here.
		 */
		_leave(" = 0 (but no root)");
		return 0;
	}

	cp = strchr(rootcell, ':');
	if (!cp) {
		printk(KERN_ERR "kAFS: no VL server IP addresses specified\n");
		_leave(" = %d (no colon)", -EINVAL);
		return -EINVAL;
	}

	/* allocate a cell record for the root cell */
	*cp++ = 0;
	ret = afs_cell_create(rootcell, cp, &new_root);
	if (ret < 0) {
		_leave(" = %d", ret);
		return ret;
	}

	/* as afs_put_cell() takes locks by itself, we have to do
	 * a little gymnastics to be race-free.
	 */
	afs_get_cell(new_root);

	write_lock(&afs_cells_lock);
	while (afs_cell_root) {
		old_root = afs_cell_root;
		afs_cell_root = NULL;
		write_unlock(&afs_cells_lock);
		afs_put_cell(old_root);
		write_lock(&afs_cells_lock);
	}
	afs_cell_root = new_root;
	write_unlock(&afs_cells_lock);

	_leave(" = %d", ret);
	return ret;
}

/*
 * lookup a cell record
 */
int afs_cell_lookup(const char *name, unsigned namesz, struct afs_cell **_cell)
{
	struct afs_cell *cell;
	int ret;

	_enter("\"%*.*s\",", namesz, namesz, name ? name : "");

	*_cell = NULL;

	if (name) {
		/* if the cell was named, look for it in the cell record list */
		ret = -ENOENT;
		cell = NULL;
		read_lock(&afs_cells_lock);

		list_for_each_entry(cell, &afs_cells, link) {
			if (strncmp(cell->name, name, namesz) == 0) {
				afs_get_cell(cell);
				goto found;
			}
		}
		cell = NULL;
	found:

		read_unlock(&afs_cells_lock);

		if (cell)
			ret = 0;
	} else {
		read_lock(&afs_cells_lock);

		cell = afs_cell_root;
		if (!cell) {
			/* this should not happen unless user tries to mount
			 * when root cell is not set. Return an impossibly
			 * bizzare errno to alert the user. Things like
			 * ENOENT might be "more appropriate" but they happen
			 * for other reasons.
			 */
			ret = -EDESTADDRREQ;
		} else {
			afs_get_cell(cell);
			ret = 0;
		}

		read_unlock(&afs_cells_lock);
	}

	*_cell = cell;
	_leave(" = %d (%p)", ret, cell);
	return ret;
}

/*
 * try and get a cell record
 */
struct afs_cell *afs_get_cell_maybe(struct afs_cell **_cell)
{
	struct afs_cell *cell;

	write_lock(&afs_cells_lock);

	cell = *_cell;
	if (cell && !list_empty(&cell->link))
		afs_get_cell(cell);
	else
		cell = NULL;

	write_unlock(&afs_cells_lock);

	return cell;
}

/*
 * destroy a cell record
 */
void afs_put_cell(struct afs_cell *cell)
{
	if (!cell)
		return;

	_enter("%p{%d,%s}", cell, atomic_read(&cell->usage), cell->name);

	/* sanity check */
	BUG_ON(atomic_read(&cell->usage) <= 0);

	/* to prevent a race, the decrement and the dequeue must be effectively
	 * atomic */
	write_lock(&afs_cells_lock);

	if (likely(!atomic_dec_and_test(&cell->usage))) {
		write_unlock(&afs_cells_lock);
		_leave("");
		return;
	}

	write_unlock(&afs_cells_lock);

	BUG_ON(!list_empty(&cell->sv_list));
	BUG_ON(!list_empty(&cell->sv_graveyard));
	BUG_ON(!list_empty(&cell->vl_list));
	BUG_ON(!list_empty(&cell->vl_graveyard));

	_leave(" [unused]");
}

/*
 * destroy a cell record
 */
static void afs_cell_destroy(struct afs_cell *cell)
{
	_enter("%p{%d,%s}", cell, atomic_read(&cell->usage), cell->name);

	/* to prevent a race, the decrement and the dequeue must be effectively
	 * atomic */
	write_lock(&afs_cells_lock);

	/* sanity check */
	BUG_ON(atomic_read(&cell->usage) != 0);

	list_del_init(&cell->link);

	write_unlock(&afs_cells_lock);

	down_write(&afs_cells_sem);

	afs_proc_cell_remove(cell);

	down_write(&afs_proc_cells_sem);
	list_del_init(&cell->proc_link);
	up_write(&afs_proc_cells_sem);

#ifdef AFS_CACHING_SUPPORT
	cachefs_relinquish_cookie(cell->cache, 0);
#endif

	up_write(&afs_cells_sem);

	BUG_ON(!list_empty(&cell->sv_list));
	BUG_ON(!list_empty(&cell->sv_graveyard));
	BUG_ON(!list_empty(&cell->vl_list));
	BUG_ON(!list_empty(&cell->vl_graveyard));

	/* finish cleaning up the cell */
	kfree(cell);

	_leave(" [destroyed]");
}

/*
 * lookup the server record corresponding to an Rx RPC peer
 */
int afs_server_find_by_peer(const struct rxrpc_peer *peer,
			    struct afs_server **_server)
{
	struct afs_server *server;
	struct afs_cell *cell;

	_enter("%p{a=%08x},", peer, ntohl(peer->addr.s_addr));

	/* search the cell list */
	read_lock(&afs_cells_lock);

	list_for_each_entry(cell, &afs_cells, link) {

		_debug("? cell %s",cell->name);

		write_lock(&cell->sv_lock);

		/* check the active list */
		list_for_each_entry(server, &cell->sv_list, link) {
			_debug("?? server %08x", ntohl(server->addr.s_addr));

			if (memcmp(&server->addr, &peer->addr,
				   sizeof(struct in_addr)) == 0)
				goto found_server;
		}

		/* check the inactive list */
		spin_lock(&cell->sv_gylock);
		list_for_each_entry(server, &cell->sv_graveyard, link) {
			_debug("?? dead server %08x",
			       ntohl(server->addr.s_addr));

			if (memcmp(&server->addr, &peer->addr,
				   sizeof(struct in_addr)) == 0)
				goto found_dead_server;
		}
		spin_unlock(&cell->sv_gylock);

		write_unlock(&cell->sv_lock);
	}
	read_unlock(&afs_cells_lock);

	_leave(" = -ENOENT");
	return -ENOENT;

	/* we found it in the graveyard - resurrect it */
found_dead_server:
	list_move_tail(&server->link, &cell->sv_list);
	afs_get_server(server);
	afs_kafstimod_del_timer(&server->timeout);
	spin_unlock(&cell->sv_gylock);
	goto success;

	/* we found it - increment its ref count and return it */
found_server:
	afs_get_server(server);

success:
	write_unlock(&cell->sv_lock);
	read_unlock(&afs_cells_lock);

	*_server = server;
	_leave(" = 0 (s=%p c=%p)", server, cell);
	return 0;
}

/*
 * purge in-memory cell database on module unload or afs_init() failure
 * - the timeout daemon is stopped before calling this
 */
void afs_cell_purge(void)
{
	struct afs_vlocation *vlocation;
	struct afs_cell *cell;

	_enter("");

	afs_put_cell(afs_cell_root);

	while (!list_empty(&afs_cells)) {
		cell = NULL;

		/* remove the next cell from the front of the list */
		write_lock(&afs_cells_lock);

		if (!list_empty(&afs_cells)) {
			cell = list_entry(afs_cells.next,
					  struct afs_cell, link);
			list_del_init(&cell->link);
		}

		write_unlock(&afs_cells_lock);

		if (cell) {
			_debug("PURGING CELL %s (%d)",
			       cell->name, atomic_read(&cell->usage));

			BUG_ON(!list_empty(&cell->sv_list));
			BUG_ON(!list_empty(&cell->vl_list));

			/* purge the cell's VL graveyard list */
			_debug(" - clearing VL graveyard");

			spin_lock(&cell->vl_gylock);

			while (!list_empty(&cell->vl_graveyard)) {
				vlocation = list_entry(cell->vl_graveyard.next,
						       struct afs_vlocation,
						       link);
				list_del_init(&vlocation->link);

				afs_kafstimod_del_timer(&vlocation->timeout);

				spin_unlock(&cell->vl_gylock);

				afs_vlocation_do_timeout(vlocation);
				/* TODO: race if move to use krxtimod instead
				 * of kafstimod */

				spin_lock(&cell->vl_gylock);
			}

			spin_unlock(&cell->vl_gylock);

			/* purge the cell's server graveyard list */
			_debug(" - clearing server graveyard");

			spin_lock(&cell->sv_gylock);

			while (!list_empty(&cell->sv_graveyard)) {
				struct afs_server *server;

				server = list_entry(cell->sv_graveyard.next,
						    struct afs_server, link);
				list_del_init(&server->link);

				afs_kafstimod_del_timer(&server->timeout);

				spin_unlock(&cell->sv_gylock);

				afs_server_do_timeout(server);

				spin_lock(&cell->sv_gylock);
			}

			spin_unlock(&cell->sv_gylock);

			/* now the cell should be left with no references */
			afs_cell_destroy(cell);
		}
	}

	_leave("");
}

/*
 * match a cell record obtained from the cache
 */
#ifdef AFS_CACHING_SUPPORT
static cachefs_match_val_t afs_cell_cache_match(void *target,
						const void *entry)
{
	const struct afs_cache_cell *ccell = entry;
	struct afs_cell *cell = target;

	_enter("{%s},{%s}", ccell->name, cell->name);

	if (strncmp(ccell->name, cell->name, sizeof(ccell->name)) == 0) {
		_leave(" = SUCCESS");
		return CACHEFS_MATCH_SUCCESS;
	}

	_leave(" = FAILED");
	return CACHEFS_MATCH_FAILED;
}
#endif

/*
 * update a cell record in the cache
 */
#ifdef AFS_CACHING_SUPPORT
static void afs_cell_cache_update(void *source, void *entry)
{
	struct afs_cache_cell *ccell = entry;
	struct afs_cell *cell = source;

	_enter("%p,%p", source, entry);

	strncpy(ccell->name, cell->name, sizeof(ccell->name));

	memcpy(ccell->vl_servers,
	       cell->vl_addrs,
	       min(sizeof(ccell->vl_servers), sizeof(cell->vl_addrs)));

}
#endif