[net-next-2.6.git] / fs / nfsd / nfssvc.c

/*
 * linux/fs/nfsd/nfssvc.c
 *
 * Central processing for nfsd.
 *
 * Authors:	Olaf Kirch (okir@monad.swb.de)
 *
 * Copyright (C) 1995, 1996, 1997 Olaf Kirch <okir@monad.swb.de>
 */

#include <linux/module.h>
#include <linux/sched.h>
#include <linux/time.h>
#include <linux/errno.h>
#include <linux/nfs.h>
#include <linux/in.h>
#include <linux/uio.h>
#include <linux/unistd.h>
#include <linux/slab.h>
#include <linux/smp.h>
#include <linux/freezer.h>
#include <linux/fs_struct.h>
#include <linux/kthread.h>
#include <linux/swap.h>

#include <linux/sunrpc/types.h>
#include <linux/sunrpc/stats.h>
#include <linux/sunrpc/svc.h>
#include <linux/sunrpc/svcsock.h>
#include <linux/sunrpc/cache.h>
#include <linux/nfsd/nfsd.h>
#include <linux/nfsd/stats.h>
#include <linux/nfsd/cache.h>
#include <linux/nfsd/syscall.h>
#include <linux/lockd/bind.h>
#include <linux/nfsacl.h>
#include <linux/seq_file.h>

#define NFSDDBG_FACILITY	NFSDDBG_SVC

extern struct svc_program	nfsd_program;
static int			nfsd(void *vrqstp);
struct timeval			nfssvc_boot;

/*
 * nfsd_mutex protects nfsd_serv -- both the pointer itself and the members
 * of the svc_serv struct. In particular, ->sv_nrthreads but also to some
 * extent ->sv_temp_socks and ->sv_permsocks. It also protects nfsdstats.th_cnt
 *
 * If (out side the lock) nfsd_serv is non-NULL, then it must point to a
 * properly initialised 'struct svc_serv' with ->sv_nrthreads > 0. That number
 * of nfsd threads must exist and each must listed in ->sp_all_threads in each
 * entry of ->sv_pools[].
 *
 * Transitions of the thread count between zero and non-zero are of particular
 * interest since the svc_serv needs to be created and initialized at that
 * point, or freed.
 *
 * Finally, the nfsd_mutex also protects some of the global variables that are
 * accessed when nfsd starts and that are settable via the write_* routines in
 * nfsctl.c. In particular:
 *
 *	user_recovery_dirname
 *	user_lease_time
 *	nfsd_versions
 */
DEFINE_MUTEX(nfsd_mutex);
struct svc_serv 		*nfsd_serv;

/*
 * nfsd_drc_lock protects nfsd_drc_max_pages and nfsd_drc_pages_used.
 * nfsd_drc_max_pages limits the total amount of memory available for
 * version 4.1 DRC caches.
 * nfsd_drc_pages_used tracks the current version 4.1 DRC memory usage.
 */
spinlock_t	nfsd_drc_lock;
unsigned int	nfsd_drc_max_mem;
unsigned int	nfsd_drc_mem_used;

#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
static struct svc_stat	nfsd_acl_svcstats;
static struct svc_version *	nfsd_acl_version[] = {
	[2] = &nfsd_acl_version2,
	[3] = &nfsd_acl_version3,
};

#define NFSD_ACL_MINVERS            2
#define NFSD_ACL_NRVERS		ARRAY_SIZE(nfsd_acl_version)
static struct svc_version *nfsd_acl_versions[NFSD_ACL_NRVERS];

static struct svc_program	nfsd_acl_program = {
	.pg_prog		= NFS_ACL_PROGRAM,
	.pg_nvers		= NFSD_ACL_NRVERS,
	.pg_vers		= nfsd_acl_versions,
	.pg_name		= "nfsacl",
	.pg_class		= "nfsd",
	.pg_stats		= &nfsd_acl_svcstats,
	.pg_authenticate	= &svc_set_client,
};

static struct svc_stat	nfsd_acl_svcstats = {
	.program	= &nfsd_acl_program,
};
#endif /* defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL) */

static struct svc_version *	nfsd_version[] = {
	[2] = &nfsd_version2,
#if defined(CONFIG_NFSD_V3)
	[3] = &nfsd_version3,
#endif
#if defined(CONFIG_NFSD_V4)
	[4] = &nfsd_version4,
#endif
};

#define NFSD_MINVERS    	2
#define NFSD_NRVERS		ARRAY_SIZE(nfsd_version)
static struct svc_version *nfsd_versions[NFSD_NRVERS];

struct svc_program		nfsd_program = {
#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
	.pg_next		= &nfsd_acl_program,
#endif
	.pg_prog		= NFS_PROGRAM,		/* program number */
	.pg_nvers		= NFSD_NRVERS,		/* nr of entries in nfsd_version */
	.pg_vers		= nfsd_versions,	/* version table */
	.pg_name		= "nfsd",		/* program name */
	.pg_class		= "nfsd",		/* authentication class */
	.pg_stats		= &nfsd_svcstats,	/* version table */
	.pg_authenticate	= &svc_set_client,	/* export authentication */

};

u32 nfsd_supported_minorversion;

int nfsd_vers(int vers, enum vers_op change)
{
	if (vers < NFSD_MINVERS || vers >= NFSD_NRVERS)
		return -1;
	switch(change) {
	case NFSD_SET:
		nfsd_versions[vers] = nfsd_version[vers];
#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
		if (vers < NFSD_ACL_NRVERS)
			nfsd_acl_versions[vers] = nfsd_acl_version[vers];
#endif
		break;
	case NFSD_CLEAR:
		nfsd_versions[vers] = NULL;
#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
		if (vers < NFSD_ACL_NRVERS)
			nfsd_acl_versions[vers] = NULL;
#endif
		break;
	case NFSD_TEST:
		return nfsd_versions[vers] != NULL;
	case NFSD_AVAIL:
		return nfsd_version[vers] != NULL;
	}
	return 0;
}

int nfsd_minorversion(u32 minorversion, enum vers_op change)
{
	if (minorversion > NFSD_SUPPORTED_MINOR_VERSION)
		return -1;
	switch(change) {
	case NFSD_SET:
		nfsd_supported_minorversion = minorversion;
		break;
	case NFSD_CLEAR:
		if (minorversion == 0)
			return -1;
		nfsd_supported_minorversion = minorversion - 1;
		break;
	case NFSD_TEST:
		return minorversion <= nfsd_supported_minorversion;
	case NFSD_AVAIL:
		return minorversion <= NFSD_SUPPORTED_MINOR_VERSION;
	}
	return 0;
}

/*
 * Maximum number of nfsd processes
 */
#define	NFSD_MAXSERVS		8192

int nfsd_nrthreads(void)
{
	int rv = 0;
	mutex_lock(&nfsd_mutex);
	if (nfsd_serv)
		rv = nfsd_serv->sv_nrthreads;
	mutex_unlock(&nfsd_mutex);
	return rv;
}

static void nfsd_last_thread(struct svc_serv *serv)
{
	/* When last nfsd thread exits we need to do some clean-up */
	struct svc_xprt *xprt;
	list_for_each_entry(xprt, &serv->sv_permsocks, xpt_list)
		lockd_down();
	nfsd_serv = NULL;
	nfsd_racache_shutdown();
	nfs4_state_shutdown();

	printk(KERN_WARNING "nfsd: last server has exited, flushing export "
			    "cache\n");
	nfsd_export_flush();
}

void nfsd_reset_versions(void)
{
	int found_one = 0;
	int i;

	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++) {
		if (nfsd_program.pg_vers[i])
			found_one = 1;
	}

	if (!found_one) {
		for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++)
			nfsd_program.pg_vers[i] = nfsd_version[i];
#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
		for (i = NFSD_ACL_MINVERS; i < NFSD_ACL_NRVERS; i++)
			nfsd_acl_program.pg_vers[i] =
				nfsd_acl_version[i];
#endif
	}
}

/*
 * Each session guarantees a negotiated per slot memory cache for replies
 * which in turn consumes memory beyond the v2/v3/v4.0 server. A dedicated
 * NFSv4.1 server might want to use more memory for a DRC than a machine
 * with mutiple services.
 *
 * Impose a hard limit on the number of pages for the DRC which varies
 * according to the machines free pages. This is of course only a default.
 *
 * For now this is a #defined shift which could be under admin control
 * in the future.
 */
static void set_max_drc(void)
{
	#define NFSD_DRC_SIZE_SHIFT	10
	nfsd_drc_max_mem = (nr_free_buffer_pages()
					>> NFSD_DRC_SIZE_SHIFT) * PAGE_SIZE;
	nfsd_drc_mem_used = 0;
	spin_lock_init(&nfsd_drc_lock);
	dprintk("%s nfsd_drc_max_mem %u \n", __func__, nfsd_drc_max_mem);
}

int nfsd_create_serv(void)
{
	int err = 0;

	WARN_ON(!mutex_is_locked(&nfsd_mutex));
	if (nfsd_serv) {
		svc_get(nfsd_serv);
		return 0;
	}
	if (nfsd_max_blksize == 0) {
		/* choose a suitable default */
		struct sysinfo i;
		si_meminfo(&i);
		/* Aim for 1/4096 of memory per thread
		 * This gives 1MB on 4Gig machines
		 * But only uses 32K on 128M machines.
		 * Bottom out at 8K on 32M and smaller.
		 * Of course, this is only a default.
		 */
		nfsd_max_blksize = NFSSVC_MAXBLKSIZE;
		i.totalram <<= PAGE_SHIFT - 12;
		while (nfsd_max_blksize > i.totalram &&
		       nfsd_max_blksize >= 8*1024*2)
			nfsd_max_blksize /= 2;
	}

	nfsd_serv = svc_create_pooled(&nfsd_program, nfsd_max_blksize,
				      nfsd_last_thread, nfsd, THIS_MODULE);
	if (nfsd_serv == NULL)
		err = -ENOMEM;
	else
		set_max_drc();

	do_gettimeofday(&nfssvc_boot);		/* record boot time */
	return err;
}

static int nfsd_init_socks(int port)
{
	int error;
	if (!list_empty(&nfsd_serv->sv_permsocks))
		return 0;

	error = svc_create_xprt(nfsd_serv, "udp", PF_INET, port,
					SVC_SOCK_DEFAULTS);
	if (error < 0)
		return error;

	error = lockd_up();
	if (error < 0)
		return error;

	error = svc_create_xprt(nfsd_serv, "tcp", PF_INET, port,
					SVC_SOCK_DEFAULTS);
	if (error < 0)
		return error;

	error = lockd_up();
	if (error < 0)
		return error;

	return 0;
}

int nfsd_nrpools(void)
{
	if (nfsd_serv == NULL)
		return 0;
	else
		return nfsd_serv->sv_nrpools;
}

int nfsd_get_nrthreads(int n, int *nthreads)
{
	int i = 0;

	if (nfsd_serv != NULL) {
		for (i = 0; i < nfsd_serv->sv_nrpools && i < n; i++)
			nthreads[i] = nfsd_serv->sv_pools[i].sp_nrthreads;
	}

	return 0;
}

int nfsd_set_nrthreads(int n, int *nthreads)
{
	int i = 0;
	int tot = 0;
	int err = 0;

	WARN_ON(!mutex_is_locked(&nfsd_mutex));

	if (nfsd_serv == NULL || n <= 0)
		return 0;

	if (n > nfsd_serv->sv_nrpools)
		n = nfsd_serv->sv_nrpools;

	/* enforce a global maximum number of threads */
	tot = 0;
	for (i = 0; i < n; i++) {
		if (nthreads[i] > NFSD_MAXSERVS)
			nthreads[i] = NFSD_MAXSERVS;
		tot += nthreads[i];
	}
	if (tot > NFSD_MAXSERVS) {
		/* total too large: scale down requested numbers */
		for (i = 0; i < n && tot > 0; i++) {
		    	int new = nthreads[i] * NFSD_MAXSERVS / tot;
			tot -= (nthreads[i] - new);
			nthreads[i] = new;
		}
		for (i = 0; i < n && tot > 0; i++) {
			nthreads[i]--;
			tot--;
		}
	}

	/*
	 * There must always be a thread in pool 0; the admin
	 * can't shut down NFS completely using pool_threads.
	 */
	if (nthreads[0] == 0)
		nthreads[0] = 1;

	/* apply the new numbers */
	svc_get(nfsd_serv);
	for (i = 0; i < n; i++) {
		err = svc_set_num_threads(nfsd_serv, &nfsd_serv->sv_pools[i],
				    	  nthreads[i]);
		if (err)
			break;
	}
	svc_destroy(nfsd_serv);

	return err;
}

int
nfsd_svc(unsigned short port, int nrservs)
{
	int	error;

	mutex_lock(&nfsd_mutex);
	dprintk("nfsd: creating service\n");
	if (nrservs <= 0)
		nrservs = 0;
	if (nrservs > NFSD_MAXSERVS)
		nrservs = NFSD_MAXSERVS;
	error = 0;
	if (nrservs == 0 && nfsd_serv == NULL)
		goto out;

	/* Readahead param cache - will no-op if it already exists */
	error =	nfsd_racache_init(2*nrservs);
	if (error<0)
		goto out;
	nfs4_state_start();

	nfsd_reset_versions();

	error = nfsd_create_serv();

	if (error)
		goto out;
	error = nfsd_init_socks(port);
	if (error)
		goto failure;

	error = svc_set_num_threads(nfsd_serv, NULL, nrservs);
	if (error == 0)
		/* We are holding a reference to nfsd_serv which
		 * we don't want to count in the return value,
		 * so subtract 1
		 */
		error = nfsd_serv->sv_nrthreads - 1;
 failure:
	svc_destroy(nfsd_serv);		/* Release server */
 out:
	mutex_unlock(&nfsd_mutex);
	return error;
}


/*
 * This is the NFS server kernel thread
 */
static int
nfsd(void *vrqstp)
{
	struct svc_rqst *rqstp = (struct svc_rqst *) vrqstp;
	int err, preverr = 0;

	/* Lock module and set up kernel thread */
	mutex_lock(&nfsd_mutex);

	/* At this point, the thread shares current->fs
	 * with the init process. We need to create files with a
	 * umask of 0 instead of init's umask. */
	if (unshare_fs_struct() < 0) {
		printk("Unable to start nfsd thread: out of memory\n");
		goto out;
	}

	current->fs->umask = 0;

	/*
	 * thread is spawned with all signals set to SIG_IGN, re-enable
	 * the ones that will bring down the thread
	 */
	allow_signal(SIGKILL);
	allow_signal(SIGHUP);
	allow_signal(SIGINT);
	allow_signal(SIGQUIT);

	nfsdstats.th_cnt++;
	mutex_unlock(&nfsd_mutex);

	/*
	 * We want less throttling in balance_dirty_pages() so that nfs to
	 * localhost doesn't cause nfsd to lock up due to all the client's
	 * dirty pages.
	 */
	current->flags |= PF_LESS_THROTTLE;
	set_freezable();

	/*
	 * The main request loop
	 */
	for (;;) {
		/*
		 * Find a socket with data available and call its
		 * recvfrom routine.
		 */
		while ((err = svc_recv(rqstp, 60*60*HZ)) == -EAGAIN)
			;
		if (err == -EINTR)
			break;
		else if (err < 0) {
			if (err != preverr) {
				printk(KERN_WARNING "%s: unexpected error "
					"from svc_recv (%d)\n", __func__, -err);
				preverr = err;
			}
			schedule_timeout_uninterruptible(HZ);
			continue;
		}


		/* Lock the export hash tables for reading. */
		exp_readlock();

		svc_process(rqstp);

		/* Unlock export hash tables */
		exp_readunlock();
	}

	/* Clear signals before calling svc_exit_thread() */
	flush_signals(current);

	mutex_lock(&nfsd_mutex);
	nfsdstats.th_cnt --;

out:
	/* Release the thread */
	svc_exit_thread(rqstp);

	/* Release module */
	mutex_unlock(&nfsd_mutex);
	module_put_and_exit(0);
	return 0;
}

static __be32 map_new_errors(u32 vers, __be32 nfserr)
{
	if (nfserr == nfserr_jukebox && vers == 2)
		return nfserr_dropit;
	if (nfserr == nfserr_wrongsec && vers < 4)
		return nfserr_acces;
	return nfserr;
}

int
nfsd_dispatch(struct svc_rqst *rqstp, __be32 *statp)
{
	struct svc_procedure	*proc;
	kxdrproc_t		xdr;
	__be32			nfserr;
	__be32			*nfserrp;

	dprintk("nfsd_dispatch: vers %d proc %d\n",
				rqstp->rq_vers, rqstp->rq_proc);
	proc = rqstp->rq_procinfo;

	/* Check whether we have this call in the cache. */
	switch (nfsd_cache_lookup(rqstp, proc->pc_cachetype)) {
	case RC_INTR:
	case RC_DROPIT:
		return 0;
	case RC_REPLY:
		return 1;
	case RC_DOIT:;
		/* do it */
	}

	/* Decode arguments */
	xdr = proc->pc_decode;
	if (xdr && !xdr(rqstp, (__be32*)rqstp->rq_arg.head[0].iov_base,
			rqstp->rq_argp)) {
		dprintk("nfsd: failed to decode arguments!\n");
		nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
		*statp = rpc_garbage_args;
		return 1;
	}

	/* need to grab the location to store the status, as
	 * nfsv4 does some encoding while processing 
	 */
	nfserrp = rqstp->rq_res.head[0].iov_base
		+ rqstp->rq_res.head[0].iov_len;
	rqstp->rq_res.head[0].iov_len += sizeof(__be32);

	/* Now call the procedure handler, and encode NFS status. */
	nfserr = proc->pc_func(rqstp, rqstp->rq_argp, rqstp->rq_resp);
	nfserr = map_new_errors(rqstp->rq_vers, nfserr);
	if (nfserr == nfserr_dropit) {
		dprintk("nfsd: Dropping request; may be revisited later\n");
		nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
		return 0;
	}

	if (rqstp->rq_proc != 0)
		*nfserrp++ = nfserr;

	/* Encode result.
	 * For NFSv2, additional info is never returned in case of an error.
	 */
	if (!(nfserr && rqstp->rq_vers == 2)) {
		xdr = proc->pc_encode;
		if (xdr && !xdr(rqstp, nfserrp,
				rqstp->rq_resp)) {
			/* Failed to encode result. Release cache entry */
			dprintk("nfsd: failed to encode result!\n");
			nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
			*statp = rpc_system_err;
			return 1;
		}
	}

	/* Store reply in cache. */
	nfsd_cache_update(rqstp, proc->pc_cachetype, statp + 1);
	return 1;
}

int nfsd_pool_stats_open(struct inode *inode, struct file *file)
{
	int ret;
	mutex_lock(&nfsd_mutex);
	if (nfsd_serv == NULL) {
		mutex_unlock(&nfsd_mutex);
		return -ENODEV;
	}
	/* bump up the psudo refcount while traversing */
	svc_get(nfsd_serv);
	ret = svc_pool_stats_open(nfsd_serv, file);
	mutex_unlock(&nfsd_mutex);
	return ret;
}

int nfsd_pool_stats_release(struct inode *inode, struct file *file)
{
	int ret = seq_release(inode, file);
	mutex_lock(&nfsd_mutex);
	/* this function really, really should have been called svc_put() */
	svc_destroy(nfsd_serv);
	mutex_unlock(&nfsd_mutex);
	return ret;
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* linux/fs/nfsd/nfssvc.c
	3	*
	4	* Central processing for nfsd.
	5	*
	6	* Authors: Olaf Kirch (okir@monad.swb.de)
	7	*
	8	* Copyright (C) 1995, 1996, 1997 Olaf Kirch <okir@monad.swb.de>
	9	*/
	10
1da177e4	11	#include <linux/module.h>
e8edc6e0	12	#include <linux/sched.h>
1da177e4 LT	13	#include <linux/time.h>
	14	#include <linux/errno.h>
	15	#include <linux/nfs.h>
	16	#include <linux/in.h>
	17	#include <linux/uio.h>
	18	#include <linux/unistd.h>
	19	#include <linux/slab.h>
	20	#include <linux/smp.h>
83144186	21	#include <linux/freezer.h>
1da177e4	22	#include <linux/fs_struct.h>
9867d76c	23	#include <linux/kthread.h>
c3d06f9c	24	#include <linux/swap.h>
1da177e4 LT	25
	26	#include <linux/sunrpc/types.h>
	27	#include <linux/sunrpc/stats.h>
	28	#include <linux/sunrpc/svc.h>
	29	#include <linux/sunrpc/svcsock.h>
	30	#include <linux/sunrpc/cache.h>
	31	#include <linux/nfsd/nfsd.h>
	32	#include <linux/nfsd/stats.h>
	33	#include <linux/nfsd/cache.h>
70c3b76c	34	#include <linux/nfsd/syscall.h>
1da177e4	35	#include <linux/lockd/bind.h>
a257cdd0	36	#include <linux/nfsacl.h>
ed2d8aed	37	#include <linux/seq_file.h>
1da177e4 LT	38
	39	#define NFSDDBG_FACILITY NFSDDBG_SVC
	40
1da177e4	41	extern struct svc_program nfsd_program;
9867d76c	42	static int nfsd(void *vrqstp);
1da177e4	43	struct timeval nfssvc_boot;
1da177e4	44
bedbdd8b NB	45	/*
	46	* nfsd_mutex protects nfsd_serv -- both the pointer itself and the members
	47	* of the svc_serv struct. In particular, ->sv_nrthreads but also to some
	48	* extent ->sv_temp_socks and ->sv_permsocks. It also protects nfsdstats.th_cnt
	49	*
	50	* If (out side the lock) nfsd_serv is non-NULL, then it must point to a
	51	* properly initialised 'struct svc_serv' with ->sv_nrthreads > 0. That number
	52	* of nfsd threads must exist and each must listed in ->sp_all_threads in each
	53	* entry of ->sv_pools[].
	54	*
	55	* Transitions of the thread count between zero and non-zero are of particular
	56	* interest since the svc_serv needs to be created and initialized at that
	57	* point, or freed.
3dd98a3b JL	58	*
	59	* Finally, the nfsd_mutex also protects some of the global variables that are
	60	* accessed when nfsd starts and that are settable via the write_* routines in
	61	* nfsctl.c. In particular:
	62	*
	63	* user_recovery_dirname
	64	* user_lease_time
	65	* nfsd_versions
bedbdd8b NB	66	*/
	67	DEFINE_MUTEX(nfsd_mutex);
	68	struct svc_serv *nfsd_serv;
	69
4bd9b0f4 AA	70	/*
	71	* nfsd_drc_lock protects nfsd_drc_max_pages and nfsd_drc_pages_used.
	72	* nfsd_drc_max_pages limits the total amount of memory available for
	73	* version 4.1 DRC caches.
	74	* nfsd_drc_pages_used tracks the current version 4.1 DRC memory usage.
	75	*/
	76	spinlock_t nfsd_drc_lock;
0c193054 AA	77	unsigned int nfsd_drc_max_mem;
0c193054 AA	78	unsigned int nfsd_drc_mem_used;
4bd9b0f4	79
3fb803a9 AG	80	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	81	static struct svc_stat nfsd_acl_svcstats;
	82	static struct svc_version * nfsd_acl_version[] = {
	83	[2] = &nfsd_acl_version2,
	84	[3] = &nfsd_acl_version3,
	85	};
	86
	87	#define NFSD_ACL_MINVERS 2
e8c96f8c	88	#define NFSD_ACL_NRVERS ARRAY_SIZE(nfsd_acl_version)
3fb803a9 AG	89	static struct svc_version *nfsd_acl_versions[NFSD_ACL_NRVERS];
	90
	91	static struct svc_program nfsd_acl_program = {
	92	.pg_prog = NFS_ACL_PROGRAM,
	93	.pg_nvers = NFSD_ACL_NRVERS,
	94	.pg_vers = nfsd_acl_versions,
1a8eff6d	95	.pg_name = "nfsacl",
3fb803a9 AG	96	.pg_class = "nfsd",
	97	.pg_stats = &nfsd_acl_svcstats,
	98	.pg_authenticate = &svc_set_client,
	99	};
	100
	101	static struct svc_stat nfsd_acl_svcstats = {
	102	.program = &nfsd_acl_program,
	103	};
	104	#endif /* defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL) */
	105
70c3b76c N	106	static struct svc_version * nfsd_version[] = {
	107	[2] = &nfsd_version2,
	108	#if defined(CONFIG_NFSD_V3)
	109	[3] = &nfsd_version3,
	110	#endif
	111	#if defined(CONFIG_NFSD_V4)
	112	[4] = &nfsd_version4,
	113	#endif
	114	};
	115
	116	#define NFSD_MINVERS 2
e8c96f8c	117	#define NFSD_NRVERS ARRAY_SIZE(nfsd_version)
70c3b76c N	118	static struct svc_version *nfsd_versions[NFSD_NRVERS];
	119
	120	struct svc_program nfsd_program = {
3fb803a9 AG	121	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	122	.pg_next = &nfsd_acl_program,
	123	#endif
70c3b76c N	124	.pg_prog = NFS_PROGRAM, /* program number */
	125	.pg_nvers = NFSD_NRVERS, /* nr of entries in nfsd_version */
	126	.pg_vers = nfsd_versions, /* version table */
	127	.pg_name = "nfsd", /* program name */
	128	.pg_class = "nfsd", /* authentication class */
	129	.pg_stats = &nfsd_svcstats, /* version table */
	130	.pg_authenticate = &svc_set_client, /* export authentication */
	131
	132	};
	133
8daf220a BH	134	u32 nfsd_supported_minorversion;
8daf220a BH	135
6658d3a7 N	136	int nfsd_vers(int vers, enum vers_op change)
	137	{
	138	if (vers < NFSD_MINVERS \|\| vers >= NFSD_NRVERS)
	139	return -1;
	140	switch(change) {
	141	case NFSD_SET:
	142	nfsd_versions[vers] = nfsd_version[vers];
6658d3a7 N	143	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
6658d3a7 N	144	if (vers < NFSD_ACL_NRVERS)
1a8eff6d	145	nfsd_acl_versions[vers] = nfsd_acl_version[vers];
6658d3a7	146	#endif
1a8eff6d	147	break;
6658d3a7 N	148	case NFSD_CLEAR:
	149	nfsd_versions[vers] = NULL;
	150	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	151	if (vers < NFSD_ACL_NRVERS)
1a8eff6d	152	nfsd_acl_versions[vers] = NULL;
6658d3a7 N	153	#endif
	154	break;
	155	case NFSD_TEST:
	156	return nfsd_versions[vers] != NULL;
	157	case NFSD_AVAIL:
	158	return nfsd_version[vers] != NULL;
	159	}
	160	return 0;
	161	}
8daf220a BH	162
	163	int nfsd_minorversion(u32 minorversion, enum vers_op change)
	164	{
	165	if (minorversion > NFSD_SUPPORTED_MINOR_VERSION)
	166	return -1;
	167	switch(change) {
	168	case NFSD_SET:
	169	nfsd_supported_minorversion = minorversion;
	170	break;
	171	case NFSD_CLEAR:
	172	if (minorversion == 0)
	173	return -1;
	174	nfsd_supported_minorversion = minorversion - 1;
	175	break;
	176	case NFSD_TEST:
	177	return minorversion <= nfsd_supported_minorversion;
	178	case NFSD_AVAIL:
	179	return minorversion <= NFSD_SUPPORTED_MINOR_VERSION;
	180	}
	181	return 0;
	182	}
	183
1da177e4 LT	184	/*
	185	* Maximum number of nfsd processes
	186	*/
	187	#define NFSD_MAXSERVS 8192
	188
	189	int nfsd_nrthreads(void)
	190	{
c7d106c9 NB	191	int rv = 0;
	192	mutex_lock(&nfsd_mutex);
	193	if (nfsd_serv)
	194	rv = nfsd_serv->sv_nrthreads;
	195	mutex_unlock(&nfsd_mutex);
	196	return rv;
1da177e4 LT	197	}
1da177e4 LT	198
bc591ccf N	199	static void nfsd_last_thread(struct svc_serv *serv)
	200	{
	201	/* When last nfsd thread exits we need to do some clean-up */
7a182083 TT	202	struct svc_xprt *xprt;
7a182083 TT	203	list_for_each_entry(xprt, &serv->sv_permsocks, xpt_list)
24e36663	204	lockd_down();
bc591ccf N	205	nfsd_serv = NULL;
	206	nfsd_racache_shutdown();
	207	nfs4_state_shutdown();
	208
e096bbc6 JL	209	printk(KERN_WARNING "nfsd: last server has exited, flushing export "
	210	"cache\n");
	211	nfsd_export_flush();
bc591ccf	212	}
6658d3a7 N	213
	214	void nfsd_reset_versions(void)
	215	{
	216	int found_one = 0;
	217	int i;
	218
	219	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++) {
	220	if (nfsd_program.pg_vers[i])
	221	found_one = 1;
	222	}
	223
	224	if (!found_one) {
	225	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++)
	226	nfsd_program.pg_vers[i] = nfsd_version[i];
	227	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	228	for (i = NFSD_ACL_MINVERS; i < NFSD_ACL_NRVERS; i++)
	229	nfsd_acl_program.pg_vers[i] =
	230	nfsd_acl_version[i];
	231	#endif
	232	}
	233	}
	234
c3d06f9c AA	235	/*
	236	* Each session guarantees a negotiated per slot memory cache for replies
	237	* which in turn consumes memory beyond the v2/v3/v4.0 server. A dedicated
	238	* NFSv4.1 server might want to use more memory for a DRC than a machine
	239	* with mutiple services.
	240	*
	241	* Impose a hard limit on the number of pages for the DRC which varies
	242	* according to the machines free pages. This is of course only a default.
	243	*
	244	* For now this is a #defined shift which could be under admin control
	245	* in the future.
	246	*/
	247	static void set_max_drc(void)
	248	{
6a14dd1a	249	#define NFSD_DRC_SIZE_SHIFT 10
0c193054 AA	250	nfsd_drc_max_mem = (nr_free_buffer_pages()
	251	>> NFSD_DRC_SIZE_SHIFT) * PAGE_SIZE;
	252	nfsd_drc_mem_used = 0;
4bd9b0f4	253	spin_lock_init(&nfsd_drc_lock);
0c193054	254	dprintk("%s nfsd_drc_max_mem %u \n", __func__, nfsd_drc_max_mem);
c3d06f9c	255	}
bedbdd8b	256
b41b66d6	257	int nfsd_create_serv(void)
02a375f0 N	258	{
02a375f0 N	259	int err = 0;
bedbdd8b NB	260
bedbdd8b NB	261	WARN_ON(!mutex_is_locked(&nfsd_mutex));
02a375f0	262	if (nfsd_serv) {
9a24ab57	263	svc_get(nfsd_serv);
02a375f0 N	264	return 0;
02a375f0 N	265	}
596bbe53 N	266	if (nfsd_max_blksize == 0) {
	267	/* choose a suitable default */
	268	struct sysinfo i;
	269	si_meminfo(&i);
	270	/* Aim for 1/4096 of memory per thread
	271	* This gives 1MB on 4Gig machines
	272	* But only uses 32K on 128M machines.
	273	* Bottom out at 8K on 32M and smaller.
	274	* Of course, this is only a default.
	275	*/
	276	nfsd_max_blksize = NFSSVC_MAXBLKSIZE;
44c55600	277	i.totalram <<= PAGE_SHIFT - 12;
596bbe53 N	278	while (nfsd_max_blksize > i.totalram &&
	279	nfsd_max_blksize >= 810242)
	280	nfsd_max_blksize /= 2;
	281	}
02a375f0	282
e096bbc6	283	nfsd_serv = svc_create_pooled(&nfsd_program, nfsd_max_blksize,
a75c5d01	284	nfsd_last_thread, nfsd, THIS_MODULE);
02a375f0 N	285	if (nfsd_serv == NULL)
02a375f0 N	286	err = -ENOMEM;
c3d06f9c AA	287	else
c3d06f9c AA	288	set_max_drc();
bedbdd8b	289
02a375f0 N	290	do_gettimeofday(&nfssvc_boot); /* record boot time */
	291	return err;
	292	}
	293
	294	static int nfsd_init_socks(int port)
	295	{
	296	int error;
	297	if (!list_empty(&nfsd_serv->sv_permsocks))
	298	return 0;
	299
9652ada3	300	error = svc_create_xprt(nfsd_serv, "udp", PF_INET, port,
482fb94e	301	SVC_SOCK_DEFAULTS);
02a375f0 N	302	if (error < 0)
	303	return error;
	304
1cd9cd16 CL	305	error = lockd_up();
	306	if (error < 0)
	307	return error;
	308
9652ada3	309	error = svc_create_xprt(nfsd_serv, "tcp", PF_INET, port,
482fb94e	310	SVC_SOCK_DEFAULTS);
02a375f0 N	311	if (error < 0)
02a375f0 N	312	return error;
26a41409 CL	313
	314	error = lockd_up();
	315	if (error < 0)
	316	return error;
	317
02a375f0 N	318	return 0;
	319	}
	320
eed2965a GB	321	int nfsd_nrpools(void)
	322	{
	323	if (nfsd_serv == NULL)
	324	return 0;
	325	else
	326	return nfsd_serv->sv_nrpools;
	327	}
	328
	329	int nfsd_get_nrthreads(int n, int *nthreads)
	330	{
	331	int i = 0;
	332
	333	if (nfsd_serv != NULL) {
	334	for (i = 0; i < nfsd_serv->sv_nrpools && i < n; i++)
	335	nthreads[i] = nfsd_serv->sv_pools[i].sp_nrthreads;
	336	}
	337
	338	return 0;
	339	}
	340
	341	int nfsd_set_nrthreads(int n, int *nthreads)
	342	{
	343	int i = 0;
	344	int tot = 0;
	345	int err = 0;
	346
bedbdd8b NB	347	WARN_ON(!mutex_is_locked(&nfsd_mutex));
bedbdd8b NB	348
eed2965a GB	349	if (nfsd_serv == NULL \|\| n <= 0)
	350	return 0;
	351
	352	if (n > nfsd_serv->sv_nrpools)
	353	n = nfsd_serv->sv_nrpools;
	354
	355	/* enforce a global maximum number of threads */
	356	tot = 0;
	357	for (i = 0; i < n; i++) {
	358	if (nthreads[i] > NFSD_MAXSERVS)
	359	nthreads[i] = NFSD_MAXSERVS;
	360	tot += nthreads[i];
	361	}
	362	if (tot > NFSD_MAXSERVS) {
	363	/* total too large: scale down requested numbers */
	364	for (i = 0; i < n && tot > 0; i++) {
	365	int new = nthreads[i] * NFSD_MAXSERVS / tot;
	366	tot -= (nthreads[i] - new);
	367	nthreads[i] = new;
	368	}
	369	for (i = 0; i < n && tot > 0; i++) {
	370	nthreads[i]--;
	371	tot--;
	372	}
	373	}
	374
	375	/*
	376	* There must always be a thread in pool 0; the admin
	377	* can't shut down NFS completely using pool_threads.
	378	*/
	379	if (nthreads[0] == 0)
	380	nthreads[0] = 1;
	381
	382	/* apply the new numbers */
eed2965a GB	383	svc_get(nfsd_serv);
	384	for (i = 0; i < n; i++) {
	385	err = svc_set_num_threads(nfsd_serv, &nfsd_serv->sv_pools[i],
	386	nthreads[i]);
	387	if (err)
	388	break;
	389	}
	390	svc_destroy(nfsd_serv);
eed2965a GB	391
	392	return err;
	393	}
	394
1da177e4 LT	395	int
	396	nfsd_svc(unsigned short port, int nrservs)
	397	{
	398	int error;
bedbdd8b NB	399
bedbdd8b NB	400	mutex_lock(&nfsd_mutex);
6658d3a7	401	dprintk("nfsd: creating service\n");
1da177e4 LT	402	if (nrservs <= 0)
	403	nrservs = 0;
	404	if (nrservs > NFSD_MAXSERVS)
	405	nrservs = NFSD_MAXSERVS;
671e1fcf N	406	error = 0;
	407	if (nrservs == 0 && nfsd_serv == NULL)
	408	goto out;
	409
1da177e4 LT	410	/* Readahead param cache - will no-op if it already exists */
	411	error = nfsd_racache_init(2*nrservs);
	412	if (error<0)
	413	goto out;
e8ff2a84	414	nfs4_state_start();
02a375f0 N	415
	416	nfsd_reset_versions();
	417
	418	error = nfsd_create_serv();
	419
	420	if (error)
	421	goto out;
	422	error = nfsd_init_socks(port);
	423	if (error)
	424	goto failure;
	425
eec09661	426	error = svc_set_num_threads(nfsd_serv, NULL, nrservs);
82e12fe9 N	427	if (error == 0)
	428	/* We are holding a reference to nfsd_serv which
	429	* we don't want to count in the return value,
	430	* so subtract 1
	431	*/
	432	error = nfsd_serv->sv_nrthreads - 1;
1da177e4	433	failure:
1da177e4	434	svc_destroy(nfsd_serv); /* Release server */
1da177e4	435	out:
bedbdd8b	436	mutex_unlock(&nfsd_mutex);
1da177e4 LT	437	return error;
	438	}
	439
1da177e4 LT	440
	441	/*
	442	* This is the NFS server kernel thread
	443	*/
9867d76c JL	444	static int
9867d76c JL	445	nfsd(void *vrqstp)
1da177e4	446	{
9867d76c	447	struct svc_rqst rqstp = (struct svc_rqst ) vrqstp;
9867d76c	448	int err, preverr = 0;
1da177e4 LT	449
1da177e4 LT	450	/* Lock module and set up kernel thread */
bedbdd8b	451	mutex_lock(&nfsd_mutex);
1da177e4	452
9867d76c	453	/* At this point, the thread shares current->fs
1da177e4 LT	454	* with the init process. We need to create files with a
1da177e4 LT	455	* umask of 0 instead of init's umask. */
3e93cd67	456	if (unshare_fs_struct() < 0) {
1da177e4 LT	457	printk("Unable to start nfsd thread: out of memory\n");
	458	goto out;
	459	}
3e93cd67	460
1da177e4 LT	461	current->fs->umask = 0;
1da177e4 LT	462
9867d76c JL	463	/*
9867d76c JL	464	* thread is spawned with all signals set to SIG_IGN, re-enable
100766f8	465	* the ones that will bring down the thread
9867d76c	466	*/
100766f8 JL	467	allow_signal(SIGKILL);
	468	allow_signal(SIGHUP);
	469	allow_signal(SIGINT);
	470	allow_signal(SIGQUIT);
bedbdd8b	471
1da177e4	472	nfsdstats.th_cnt++;
bedbdd8b NB	473	mutex_unlock(&nfsd_mutex);
bedbdd8b NB	474
1da177e4 LT	475	/*
	476	* We want less throttling in balance_dirty_pages() so that nfs to
	477	* localhost doesn't cause nfsd to lock up due to all the client's
	478	* dirty pages.
	479	*/
	480	current->flags \|= PF_LESS_THROTTLE;
83144186	481	set_freezable();
1da177e4 LT	482
	483	/*
	484	* The main request loop
	485	*/
	486	for (;;) {
1da177e4 LT	487	/*
	488	* Find a socket with data available and call its
	489	* recvfrom routine.
	490	*/
6fb2b47f	491	while ((err = svc_recv(rqstp, 6060HZ)) == -EAGAIN)
1da177e4	492	;
9867d76c	493	if (err == -EINTR)
1da177e4	494	break;
9867d76c JL	495	else if (err < 0) {
	496	if (err != preverr) {
	497	printk(KERN_WARNING "%s: unexpected error "
	498	"from svc_recv (%d)\n", __func__, -err);
	499	preverr = err;
	500	}
	501	schedule_timeout_uninterruptible(HZ);
	502	continue;
	503	}
	504
1da177e4 LT	505
	506	/* Lock the export hash tables for reading. */
	507	exp_readlock();
	508
6fb2b47f	509	svc_process(rqstp);
1da177e4 LT	510
	511	/* Unlock export hash tables */
	512	exp_readunlock();
1da177e4 LT	513	}
1da177e4 LT	514
24e36663	515	/* Clear signals before calling svc_exit_thread() */
9e416052	516	flush_signals(current);
1da177e4	517
bedbdd8b	518	mutex_lock(&nfsd_mutex);
1da177e4 LT	519	nfsdstats.th_cnt --;
	520
	521	out:
	522	/* Release the thread */
	523	svc_exit_thread(rqstp);
	524
	525	/* Release module */
bedbdd8b	526	mutex_unlock(&nfsd_mutex);
1da177e4	527	module_put_and_exit(0);
9867d76c	528	return 0;
1da177e4 LT	529	}
1da177e4 LT	530
32c1eb0c AA	531	static __be32 map_new_errors(u32 vers, __be32 nfserr)
	532	{
	533	if (nfserr == nfserr_jukebox && vers == 2)
	534	return nfserr_dropit;
	535	if (nfserr == nfserr_wrongsec && vers < 4)
	536	return nfserr_acces;
	537	return nfserr;
	538	}
	539
1da177e4	540	int
c7afef1f	541	nfsd_dispatch(struct svc_rqst rqstp, __be32 statp)
1da177e4 LT	542	{
	543	struct svc_procedure *proc;
	544	kxdrproc_t xdr;
ad451d38 AV	545	__be32 nfserr;
ad451d38 AV	546	__be32 *nfserrp;
1da177e4 LT	547
	548	dprintk("nfsd_dispatch: vers %d proc %d\n",
	549	rqstp->rq_vers, rqstp->rq_proc);
	550	proc = rqstp->rq_procinfo;
	551
	552	/* Check whether we have this call in the cache. */
	553	switch (nfsd_cache_lookup(rqstp, proc->pc_cachetype)) {
	554	case RC_INTR:
	555	case RC_DROPIT:
	556	return 0;
	557	case RC_REPLY:
	558	return 1;
	559	case RC_DOIT:;
	560	/* do it */
	561	}
	562
	563	/* Decode arguments */
	564	xdr = proc->pc_decode;
ad451d38	565	if (xdr && !xdr(rqstp, (__be32*)rqstp->rq_arg.head[0].iov_base,
1da177e4 LT	566	rqstp->rq_argp)) {
	567	dprintk("nfsd: failed to decode arguments!\n");
	568	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	569	*statp = rpc_garbage_args;
	570	return 1;
	571	}
	572
	573	/* need to grab the location to store the status, as
	574	* nfsv4 does some encoding while processing
	575	*/
	576	nfserrp = rqstp->rq_res.head[0].iov_base
	577	+ rqstp->rq_res.head[0].iov_len;
ad451d38	578	rqstp->rq_res.head[0].iov_len += sizeof(__be32);
1da177e4 LT	579
	580	/* Now call the procedure handler, and encode NFS status. */
	581	nfserr = proc->pc_func(rqstp, rqstp->rq_argp, rqstp->rq_resp);
32c1eb0c	582	nfserr = map_new_errors(rqstp->rq_vers, nfserr);
1da177e4	583	if (nfserr == nfserr_dropit) {
45457e09	584	dprintk("nfsd: Dropping request; may be revisited later\n");
1da177e4 LT	585	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	586	return 0;
	587	}
	588
	589	if (rqstp->rq_proc != 0)
	590	*nfserrp++ = nfserr;
	591
	592	/* Encode result.
	593	* For NFSv2, additional info is never returned in case of an error.
	594	*/
	595	if (!(nfserr && rqstp->rq_vers == 2)) {
	596	xdr = proc->pc_encode;
	597	if (xdr && !xdr(rqstp, nfserrp,
	598	rqstp->rq_resp)) {
	599	/* Failed to encode result. Release cache entry */
	600	dprintk("nfsd: failed to encode result!\n");
	601	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	602	*statp = rpc_system_err;
	603	return 1;
	604	}
	605	}
	606
	607	/* Store reply in cache. */
	608	nfsd_cache_update(rqstp, proc->pc_cachetype, statp + 1);
	609	return 1;
	610	}
03cf6c9f GB	611
	612	int nfsd_pool_stats_open(struct inode inode, struct file file)
	613	{
ed2d8aed RY	614	int ret;
	615	mutex_lock(&nfsd_mutex);
	616	if (nfsd_serv == NULL) {
	617	mutex_unlock(&nfsd_mutex);
03cf6c9f	618	return -ENODEV;
ed2d8aed RY	619	}
	620	/* bump up the psudo refcount while traversing */
	621	svc_get(nfsd_serv);
	622	ret = svc_pool_stats_open(nfsd_serv, file);
	623	mutex_unlock(&nfsd_mutex);
	624	return ret;
	625	}
	626
	627	int nfsd_pool_stats_release(struct inode inode, struct file file)
	628	{
	629	int ret = seq_release(inode, file);
	630	mutex_lock(&nfsd_mutex);
	631	/* this function really, really should have been called svc_put() */
	632	svc_destroy(nfsd_serv);
	633	mutex_unlock(&nfsd_mutex);
	634	return ret;
03cf6c9f	635	}