[linux.git] / fs / nfsd / nfssvc.c

/*
 * linux/fs/nfsd/nfssvc.c
 *
 * Central processing for nfsd.
 *
 * Authors:	Olaf Kirch ([email protected])
 *
 * Copyright (C) 1995, 1996, 1997 Olaf Kirch <[email protected]>
 */

#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/smp_lock.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>

#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;

#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)
{
	/* The percent of nr_free_buffer_pages used by the V4.1 server DRC */
	#define NFSD_DRC_SIZE_SHIFT	7
	nfsd_serv->sv_drc_max_pages = nr_free_buffer_pages()
						>> NFSD_DRC_SIZE_SHIFT;
	nfsd_serv->sv_drc_pages_used = 0;
	dprintk("%s svc_drc_max_pages %u\n", __func__,
		nfsd_serv->sv_drc_max_pages);
}

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");
	error = -EINVAL;
	if (nrservs <= 0)
		nrservs = 0;
	if (nrservs > NFSD_MAXSERVS)
		nrservs = NFSD_MAXSERVS;
	
	/* 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);
 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);

	/* NFSv4.1 DRC requires statp */
	if (rqstp->rq_vers == 4)
		nfsd4_set_statp(rqstp, statp);

	/* 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)
{
	if (nfsd_serv == NULL)
		return -ENODEV;
	return svc_pool_stats_open(nfsd_serv, file);
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* linux/fs/nfsd/nfssvc.c
	3	*
	4	* Central processing for nfsd.
	5	*
	6	* Authors: Olaf Kirch ([email protected])
	7	*
	8	* Copyright (C) 1995, 1996, 1997 Olaf Kirch <[email protected]>
	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>
	21	#include <linux/smp_lock.h>
83144186	22	#include <linux/freezer.h>
1da177e4	23	#include <linux/fs_struct.h>
9867d76c	24	#include <linux/kthread.h>
c3d06f9c	25	#include <linux/swap.h>
1da177e4 LT	26
	27	#include <linux/sunrpc/types.h>
	28	#include <linux/sunrpc/stats.h>
	29	#include <linux/sunrpc/svc.h>
	30	#include <linux/sunrpc/svcsock.h>
	31	#include <linux/sunrpc/cache.h>
	32	#include <linux/nfsd/nfsd.h>
	33	#include <linux/nfsd/stats.h>
	34	#include <linux/nfsd/cache.h>
70c3b76c	35	#include <linux/nfsd/syscall.h>
1da177e4	36	#include <linux/lockd/bind.h>
a257cdd0	37	#include <linux/nfsacl.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
3fb803a9 AG	70	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	71	static struct svc_stat nfsd_acl_svcstats;
	72	static struct svc_version * nfsd_acl_version[] = {
	73	[2] = &nfsd_acl_version2,
	74	[3] = &nfsd_acl_version3,
	75	};
	76
	77	#define NFSD_ACL_MINVERS 2
e8c96f8c	78	#define NFSD_ACL_NRVERS ARRAY_SIZE(nfsd_acl_version)
3fb803a9 AG	79	static struct svc_version *nfsd_acl_versions[NFSD_ACL_NRVERS];
	80
	81	static struct svc_program nfsd_acl_program = {
	82	.pg_prog = NFS_ACL_PROGRAM,
	83	.pg_nvers = NFSD_ACL_NRVERS,
	84	.pg_vers = nfsd_acl_versions,
1a8eff6d	85	.pg_name = "nfsacl",
3fb803a9 AG	86	.pg_class = "nfsd",
	87	.pg_stats = &nfsd_acl_svcstats,
	88	.pg_authenticate = &svc_set_client,
	89	};
	90
	91	static struct svc_stat nfsd_acl_svcstats = {
	92	.program = &nfsd_acl_program,
	93	};
	94	#endif /* defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL) */
	95
70c3b76c N	96	static struct svc_version * nfsd_version[] = {
	97	[2] = &nfsd_version2,
	98	#if defined(CONFIG_NFSD_V3)
	99	[3] = &nfsd_version3,
	100	#endif
	101	#if defined(CONFIG_NFSD_V4)
	102	[4] = &nfsd_version4,
	103	#endif
	104	};
	105
	106	#define NFSD_MINVERS 2
e8c96f8c	107	#define NFSD_NRVERS ARRAY_SIZE(nfsd_version)
70c3b76c N	108	static struct svc_version *nfsd_versions[NFSD_NRVERS];
	109
	110	struct svc_program nfsd_program = {
3fb803a9 AG	111	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	112	.pg_next = &nfsd_acl_program,
	113	#endif
70c3b76c N	114	.pg_prog = NFS_PROGRAM, /* program number */
	115	.pg_nvers = NFSD_NRVERS, /* nr of entries in nfsd_version */
	116	.pg_vers = nfsd_versions, /* version table */
	117	.pg_name = "nfsd", /* program name */
	118	.pg_class = "nfsd", /* authentication class */
	119	.pg_stats = &nfsd_svcstats, /* version table */
	120	.pg_authenticate = &svc_set_client, /* export authentication */
	121
	122	};
	123
8daf220a BH	124	u32 nfsd_supported_minorversion;
8daf220a BH	125
6658d3a7 N	126	int nfsd_vers(int vers, enum vers_op change)
	127	{
	128	if (vers < NFSD_MINVERS \|\| vers >= NFSD_NRVERS)
	129	return -1;
	130	switch(change) {
	131	case NFSD_SET:
	132	nfsd_versions[vers] = nfsd_version[vers];
6658d3a7 N	133	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
6658d3a7 N	134	if (vers < NFSD_ACL_NRVERS)
1a8eff6d	135	nfsd_acl_versions[vers] = nfsd_acl_version[vers];
6658d3a7	136	#endif
1a8eff6d	137	break;
6658d3a7 N	138	case NFSD_CLEAR:
	139	nfsd_versions[vers] = NULL;
	140	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	141	if (vers < NFSD_ACL_NRVERS)
1a8eff6d	142	nfsd_acl_versions[vers] = NULL;
6658d3a7 N	143	#endif
	144	break;
	145	case NFSD_TEST:
	146	return nfsd_versions[vers] != NULL;
	147	case NFSD_AVAIL:
	148	return nfsd_version[vers] != NULL;
	149	}
	150	return 0;
	151	}
8daf220a BH	152
	153	int nfsd_minorversion(u32 minorversion, enum vers_op change)
	154	{
	155	if (minorversion > NFSD_SUPPORTED_MINOR_VERSION)
	156	return -1;
	157	switch(change) {
	158	case NFSD_SET:
	159	nfsd_supported_minorversion = minorversion;
	160	break;
	161	case NFSD_CLEAR:
	162	if (minorversion == 0)
	163	return -1;
	164	nfsd_supported_minorversion = minorversion - 1;
	165	break;
	166	case NFSD_TEST:
	167	return minorversion <= nfsd_supported_minorversion;
	168	case NFSD_AVAIL:
	169	return minorversion <= NFSD_SUPPORTED_MINOR_VERSION;
	170	}
	171	return 0;
	172	}
	173
1da177e4 LT	174	/*
	175	* Maximum number of nfsd processes
	176	*/
	177	#define NFSD_MAXSERVS 8192
	178
	179	int nfsd_nrthreads(void)
	180	{
c7d106c9 NB	181	int rv = 0;
	182	mutex_lock(&nfsd_mutex);
	183	if (nfsd_serv)
	184	rv = nfsd_serv->sv_nrthreads;
	185	mutex_unlock(&nfsd_mutex);
	186	return rv;
1da177e4 LT	187	}
1da177e4 LT	188
bc591ccf N	189	static void nfsd_last_thread(struct svc_serv *serv)
	190	{
	191	/* When last nfsd thread exits we need to do some clean-up */
7a182083 TT	192	struct svc_xprt *xprt;
7a182083 TT	193	list_for_each_entry(xprt, &serv->sv_permsocks, xpt_list)
24e36663	194	lockd_down();
bc591ccf N	195	nfsd_serv = NULL;
	196	nfsd_racache_shutdown();
	197	nfs4_state_shutdown();
	198
e096bbc6 JL	199	printk(KERN_WARNING "nfsd: last server has exited, flushing export "
	200	"cache\n");
	201	nfsd_export_flush();
bc591ccf	202	}
6658d3a7 N	203
	204	void nfsd_reset_versions(void)
	205	{
	206	int found_one = 0;
	207	int i;
	208
	209	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++) {
	210	if (nfsd_program.pg_vers[i])
	211	found_one = 1;
	212	}
	213
	214	if (!found_one) {
	215	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++)
	216	nfsd_program.pg_vers[i] = nfsd_version[i];
	217	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	218	for (i = NFSD_ACL_MINVERS; i < NFSD_ACL_NRVERS; i++)
	219	nfsd_acl_program.pg_vers[i] =
	220	nfsd_acl_version[i];
	221	#endif
	222	}
	223	}
	224
c3d06f9c AA	225	/*
	226	* Each session guarantees a negotiated per slot memory cache for replies
	227	* which in turn consumes memory beyond the v2/v3/v4.0 server. A dedicated
	228	* NFSv4.1 server might want to use more memory for a DRC than a machine
	229	* with mutiple services.
	230	*
	231	* Impose a hard limit on the number of pages for the DRC which varies
	232	* according to the machines free pages. This is of course only a default.
	233	*
	234	* For now this is a #defined shift which could be under admin control
	235	* in the future.
	236	*/
	237	static void set_max_drc(void)
	238	{
f0ad670d BH	239	/* The percent of nr_free_buffer_pages used by the V4.1 server DRC */
f0ad670d BH	240	#define NFSD_DRC_SIZE_SHIFT 7
c3d06f9c AA	241	nfsd_serv->sv_drc_max_pages = nr_free_buffer_pages()
	242	>> NFSD_DRC_SIZE_SHIFT;
	243	nfsd_serv->sv_drc_pages_used = 0;
	244	dprintk("%s svc_drc_max_pages %u\n", __func__,
	245	nfsd_serv->sv_drc_max_pages);
	246	}
bedbdd8b	247
b41b66d6	248	int nfsd_create_serv(void)
02a375f0 N	249	{
02a375f0 N	250	int err = 0;
bedbdd8b NB	251
bedbdd8b NB	252	WARN_ON(!mutex_is_locked(&nfsd_mutex));
02a375f0	253	if (nfsd_serv) {
9a24ab57	254	svc_get(nfsd_serv);
02a375f0 N	255	return 0;
02a375f0 N	256	}
596bbe53 N	257	if (nfsd_max_blksize == 0) {
	258	/* choose a suitable default */
	259	struct sysinfo i;
	260	si_meminfo(&i);
	261	/* Aim for 1/4096 of memory per thread
	262	* This gives 1MB on 4Gig machines
	263	* But only uses 32K on 128M machines.
	264	* Bottom out at 8K on 32M and smaller.
	265	* Of course, this is only a default.
	266	*/
	267	nfsd_max_blksize = NFSSVC_MAXBLKSIZE;
44c55600	268	i.totalram <<= PAGE_SHIFT - 12;
596bbe53 N	269	while (nfsd_max_blksize > i.totalram &&
	270	nfsd_max_blksize >= 810242)
	271	nfsd_max_blksize /= 2;
	272	}
02a375f0	273
e096bbc6	274	nfsd_serv = svc_create_pooled(&nfsd_program, nfsd_max_blksize,
a75c5d01	275	nfsd_last_thread, nfsd, THIS_MODULE);
02a375f0 N	276	if (nfsd_serv == NULL)
02a375f0 N	277	err = -ENOMEM;
c3d06f9c AA	278	else
c3d06f9c AA	279	set_max_drc();
bedbdd8b	280
02a375f0 N	281	do_gettimeofday(&nfssvc_boot); /* record boot time */
	282	return err;
	283	}
	284
	285	static int nfsd_init_socks(int port)
	286	{
	287	int error;
	288	if (!list_empty(&nfsd_serv->sv_permsocks))
	289	return 0;
	290
9652ada3	291	error = svc_create_xprt(nfsd_serv, "udp", PF_INET, port,
482fb94e	292	SVC_SOCK_DEFAULTS);
02a375f0 N	293	if (error < 0)
	294	return error;
	295
1cd9cd16 CL	296	error = lockd_up();
	297	if (error < 0)
	298	return error;
	299
9652ada3	300	error = svc_create_xprt(nfsd_serv, "tcp", PF_INET, port,
482fb94e	301	SVC_SOCK_DEFAULTS);
02a375f0 N	302	if (error < 0)
02a375f0 N	303	return error;
26a41409 CL	304
	305	error = lockd_up();
	306	if (error < 0)
	307	return error;
	308
02a375f0 N	309	return 0;
	310	}
	311
eed2965a GB	312	int nfsd_nrpools(void)
	313	{
	314	if (nfsd_serv == NULL)
	315	return 0;
	316	else
	317	return nfsd_serv->sv_nrpools;
	318	}
	319
	320	int nfsd_get_nrthreads(int n, int *nthreads)
	321	{
	322	int i = 0;
	323
	324	if (nfsd_serv != NULL) {
	325	for (i = 0; i < nfsd_serv->sv_nrpools && i < n; i++)
	326	nthreads[i] = nfsd_serv->sv_pools[i].sp_nrthreads;
	327	}
	328
	329	return 0;
	330	}
	331
	332	int nfsd_set_nrthreads(int n, int *nthreads)
	333	{
	334	int i = 0;
	335	int tot = 0;
	336	int err = 0;
	337
bedbdd8b NB	338	WARN_ON(!mutex_is_locked(&nfsd_mutex));
bedbdd8b NB	339
eed2965a GB	340	if (nfsd_serv == NULL \|\| n <= 0)
	341	return 0;
	342
	343	if (n > nfsd_serv->sv_nrpools)
	344	n = nfsd_serv->sv_nrpools;
	345
	346	/* enforce a global maximum number of threads */
	347	tot = 0;
	348	for (i = 0; i < n; i++) {
	349	if (nthreads[i] > NFSD_MAXSERVS)
	350	nthreads[i] = NFSD_MAXSERVS;
	351	tot += nthreads[i];
	352	}
	353	if (tot > NFSD_MAXSERVS) {
	354	/* total too large: scale down requested numbers */
	355	for (i = 0; i < n && tot > 0; i++) {
	356	int new = nthreads[i] * NFSD_MAXSERVS / tot;
	357	tot -= (nthreads[i] - new);
	358	nthreads[i] = new;
	359	}
	360	for (i = 0; i < n && tot > 0; i++) {
	361	nthreads[i]--;
	362	tot--;
	363	}
	364	}
	365
	366	/*
	367	* There must always be a thread in pool 0; the admin
	368	* can't shut down NFS completely using pool_threads.
	369	*/
	370	if (nthreads[0] == 0)
	371	nthreads[0] = 1;
	372
	373	/* apply the new numbers */
eed2965a GB	374	svc_get(nfsd_serv);
	375	for (i = 0; i < n; i++) {
	376	err = svc_set_num_threads(nfsd_serv, &nfsd_serv->sv_pools[i],
	377	nthreads[i]);
	378	if (err)
	379	break;
	380	}
	381	svc_destroy(nfsd_serv);
eed2965a GB	382
	383	return err;
	384	}
	385
1da177e4 LT	386	int
	387	nfsd_svc(unsigned short port, int nrservs)
	388	{
	389	int error;
bedbdd8b NB	390
bedbdd8b NB	391	mutex_lock(&nfsd_mutex);
6658d3a7	392	dprintk("nfsd: creating service\n");
1da177e4 LT	393	error = -EINVAL;
	394	if (nrservs <= 0)
	395	nrservs = 0;
	396	if (nrservs > NFSD_MAXSERVS)
	397	nrservs = NFSD_MAXSERVS;
	398
	399	/* Readahead param cache - will no-op if it already exists */
	400	error = nfsd_racache_init(2*nrservs);
	401	if (error<0)
	402	goto out;
e8ff2a84	403	nfs4_state_start();
02a375f0 N	404
	405	nfsd_reset_versions();
	406
	407	error = nfsd_create_serv();
	408
	409	if (error)
	410	goto out;
	411	error = nfsd_init_socks(port);
	412	if (error)
	413	goto failure;
	414
eec09661	415	error = svc_set_num_threads(nfsd_serv, NULL, nrservs);
1da177e4	416	failure:
1da177e4	417	svc_destroy(nfsd_serv); /* Release server */
1da177e4	418	out:
bedbdd8b	419	mutex_unlock(&nfsd_mutex);
1da177e4 LT	420	return error;
	421	}
	422
1da177e4 LT	423
	424	/*
	425	* This is the NFS server kernel thread
	426	*/
9867d76c JL	427	static int
9867d76c JL	428	nfsd(void *vrqstp)
1da177e4	429	{
9867d76c	430	struct svc_rqst rqstp = (struct svc_rqst ) vrqstp;
9867d76c	431	int err, preverr = 0;
1da177e4 LT	432
1da177e4 LT	433	/* Lock module and set up kernel thread */
bedbdd8b	434	mutex_lock(&nfsd_mutex);
1da177e4	435
9867d76c	436	/* At this point, the thread shares current->fs
1da177e4 LT	437	* with the init process. We need to create files with a
1da177e4 LT	438	* umask of 0 instead of init's umask. */
3e93cd67	439	if (unshare_fs_struct() < 0) {
1da177e4 LT	440	printk("Unable to start nfsd thread: out of memory\n");
	441	goto out;
	442	}
3e93cd67	443
1da177e4 LT	444	current->fs->umask = 0;
1da177e4 LT	445
9867d76c JL	446	/*
9867d76c JL	447	* thread is spawned with all signals set to SIG_IGN, re-enable
100766f8	448	* the ones that will bring down the thread
9867d76c	449	*/
100766f8 JL	450	allow_signal(SIGKILL);
	451	allow_signal(SIGHUP);
	452	allow_signal(SIGINT);
	453	allow_signal(SIGQUIT);
bedbdd8b	454
1da177e4	455	nfsdstats.th_cnt++;
bedbdd8b NB	456	mutex_unlock(&nfsd_mutex);
bedbdd8b NB	457
1da177e4 LT	458	/*
	459	* We want less throttling in balance_dirty_pages() so that nfs to
	460	* localhost doesn't cause nfsd to lock up due to all the client's
	461	* dirty pages.
	462	*/
	463	current->flags \|= PF_LESS_THROTTLE;
83144186	464	set_freezable();
1da177e4 LT	465
	466	/*
	467	* The main request loop
	468	*/
	469	for (;;) {
1da177e4 LT	470	/*
	471	* Find a socket with data available and call its
	472	* recvfrom routine.
	473	*/
6fb2b47f	474	while ((err = svc_recv(rqstp, 6060HZ)) == -EAGAIN)
1da177e4	475	;
9867d76c	476	if (err == -EINTR)
1da177e4	477	break;
9867d76c JL	478	else if (err < 0) {
	479	if (err != preverr) {
	480	printk(KERN_WARNING "%s: unexpected error "
	481	"from svc_recv (%d)\n", __func__, -err);
	482	preverr = err;
	483	}
	484	schedule_timeout_uninterruptible(HZ);
	485	continue;
	486	}
	487
1da177e4 LT	488
	489	/* Lock the export hash tables for reading. */
	490	exp_readlock();
	491
6fb2b47f	492	svc_process(rqstp);
1da177e4 LT	493
	494	/* Unlock export hash tables */
	495	exp_readunlock();
1da177e4 LT	496	}
1da177e4 LT	497
24e36663	498	/* Clear signals before calling svc_exit_thread() */
9e416052	499	flush_signals(current);
1da177e4	500
bedbdd8b	501	mutex_lock(&nfsd_mutex);
1da177e4 LT	502	nfsdstats.th_cnt --;
	503
	504	out:
	505	/* Release the thread */
	506	svc_exit_thread(rqstp);
	507
	508	/* Release module */
bedbdd8b	509	mutex_unlock(&nfsd_mutex);
1da177e4	510	module_put_and_exit(0);
9867d76c	511	return 0;
1da177e4 LT	512	}
1da177e4 LT	513
32c1eb0c AA	514	static __be32 map_new_errors(u32 vers, __be32 nfserr)
	515	{
	516	if (nfserr == nfserr_jukebox && vers == 2)
	517	return nfserr_dropit;
	518	if (nfserr == nfserr_wrongsec && vers < 4)
	519	return nfserr_acces;
	520	return nfserr;
	521	}
	522
1da177e4	523	int
c7afef1f	524	nfsd_dispatch(struct svc_rqst rqstp, __be32 statp)
1da177e4 LT	525	{
	526	struct svc_procedure *proc;
	527	kxdrproc_t xdr;
ad451d38 AV	528	__be32 nfserr;
ad451d38 AV	529	__be32 *nfserrp;
1da177e4 LT	530
	531	dprintk("nfsd_dispatch: vers %d proc %d\n",
	532	rqstp->rq_vers, rqstp->rq_proc);
	533	proc = rqstp->rq_procinfo;
	534
	535	/* Check whether we have this call in the cache. */
	536	switch (nfsd_cache_lookup(rqstp, proc->pc_cachetype)) {
	537	case RC_INTR:
	538	case RC_DROPIT:
	539	return 0;
	540	case RC_REPLY:
	541	return 1;
	542	case RC_DOIT:;
	543	/* do it */
	544	}
	545
	546	/* Decode arguments */
	547	xdr = proc->pc_decode;
ad451d38	548	if (xdr && !xdr(rqstp, (__be32*)rqstp->rq_arg.head[0].iov_base,
1da177e4 LT	549	rqstp->rq_argp)) {
	550	dprintk("nfsd: failed to decode arguments!\n");
	551	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	552	*statp = rpc_garbage_args;
	553	return 1;
	554	}
	555
	556	/* need to grab the location to store the status, as
	557	* nfsv4 does some encoding while processing
	558	*/
	559	nfserrp = rqstp->rq_res.head[0].iov_base
	560	+ rqstp->rq_res.head[0].iov_len;
ad451d38	561	rqstp->rq_res.head[0].iov_len += sizeof(__be32);
1da177e4	562
074fe897 AA	563	/* NFSv4.1 DRC requires statp */
	564	if (rqstp->rq_vers == 4)
	565	nfsd4_set_statp(rqstp, statp);
	566
1da177e4 LT	567	/* Now call the procedure handler, and encode NFS status. */
1da177e4 LT	568	nfserr = proc->pc_func(rqstp, rqstp->rq_argp, rqstp->rq_resp);
32c1eb0c	569	nfserr = map_new_errors(rqstp->rq_vers, nfserr);
1da177e4	570	if (nfserr == nfserr_dropit) {
45457e09	571	dprintk("nfsd: Dropping request; may be revisited later\n");
1da177e4 LT	572	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	573	return 0;
	574	}
	575
	576	if (rqstp->rq_proc != 0)
	577	*nfserrp++ = nfserr;
	578
	579	/* Encode result.
	580	* For NFSv2, additional info is never returned in case of an error.
	581	*/
	582	if (!(nfserr && rqstp->rq_vers == 2)) {
	583	xdr = proc->pc_encode;
	584	if (xdr && !xdr(rqstp, nfserrp,
	585	rqstp->rq_resp)) {
	586	/* Failed to encode result. Release cache entry */
	587	dprintk("nfsd: failed to encode result!\n");
	588	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	589	*statp = rpc_system_err;
	590	return 1;
	591	}
	592	}
	593
	594	/* Store reply in cache. */
	595	nfsd_cache_update(rqstp, proc->pc_cachetype, statp + 1);
	596	return 1;
	597	}
03cf6c9f GB	598
	599	int nfsd_pool_stats_open(struct inode inode, struct file file)
	600	{
	601	if (nfsd_serv == NULL)
	602	return -ENODEV;
	603	return svc_pool_stats_open(nfsd_serv, file);
	604	}