[linux-2.6-block.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/smp_lock.h>
#include <linux/freezer.h>
#include <linux/fs_struct.h>
#include <linux/kthread.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;
static atomic_t			nfsd_busy;
static unsigned long		nfsd_last_call;
static DEFINE_SPINLOCK(nfsd_call_lock);

/*
 * 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 */

};

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;
}
/*
 * 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
	}
}


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;
	}

	atomic_set(&nfsd_busy, 0);
	nfsd_serv = svc_create_pooled(&nfsd_program, nfsd_max_blksize,
				      AF_INET,
				      nfsd_last_thread, nfsd, THIS_MODULE);
	if (nfsd_serv == NULL)
		err = -ENOMEM;

	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", port,
					SVC_SOCK_DEFAULTS);
	if (error < 0)
		return error;

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

	error = svc_create_xprt(nfsd_serv, "tcp", 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;
}

static inline void
update_thread_usage(int busy_threads)
{
	unsigned long prev_call;
	unsigned long diff;
	int decile;

	spin_lock(&nfsd_call_lock);
	prev_call = nfsd_last_call;
	nfsd_last_call = jiffies;
	decile = busy_threads*10/nfsdstats.th_cnt;
	if (decile>0 && decile <= 10) {
		diff = nfsd_last_call - prev_call;
		if ( (nfsdstats.th_usage[decile-1] += diff) >= NFSD_USAGE_WRAP)
			nfsdstats.th_usage[decile-1] -= NFSD_USAGE_WRAP;
		if (decile == 10)
			nfsdstats.th_fullcnt++;
	}
	spin_unlock(&nfsd_call_lock);
}

/*
 * 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;
		}

		update_thread_usage(atomic_read(&nfsd_busy));
		atomic_inc(&nfsd_busy);

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

		svc_process(rqstp);

		/* Unlock export hash tables */
		exp_readunlock();
		update_thread_usage(atomic_read(&nfsd_busy));
		atomic_dec(&nfsd_busy);
	}

	/* 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;
}
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>
	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>
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>
1da177e4 LT	37
	38	#define NFSDDBG_FACILITY NFSDDBG_SVC
	39
1da177e4	40	extern struct svc_program nfsd_program;
9867d76c	41	static int nfsd(void *vrqstp);
1da177e4	42	struct timeval nfssvc_boot;
1da177e4 LT	43	static atomic_t nfsd_busy;
	44	static unsigned long nfsd_last_call;
	45	static DEFINE_SPINLOCK(nfsd_call_lock);
	46
bedbdd8b NB	47	/*
	48	* nfsd_mutex protects nfsd_serv -- both the pointer itself and the members
	49	* of the svc_serv struct. In particular, ->sv_nrthreads but also to some
	50	* extent ->sv_temp_socks and ->sv_permsocks. It also protects nfsdstats.th_cnt
	51	*
	52	* If (out side the lock) nfsd_serv is non-NULL, then it must point to a
	53	* properly initialised 'struct svc_serv' with ->sv_nrthreads > 0. That number
	54	* of nfsd threads must exist and each must listed in ->sp_all_threads in each
	55	* entry of ->sv_pools[].
	56	*
	57	* Transitions of the thread count between zero and non-zero are of particular
	58	* interest since the svc_serv needs to be created and initialized at that
	59	* point, or freed.
3dd98a3b JL	60	*
	61	* Finally, the nfsd_mutex also protects some of the global variables that are
	62	* accessed when nfsd starts and that are settable via the write_* routines in
	63	* nfsctl.c. In particular:
	64	*
	65	* user_recovery_dirname
	66	* user_lease_time
	67	* nfsd_versions
bedbdd8b NB	68	*/
	69	DEFINE_MUTEX(nfsd_mutex);
	70	struct svc_serv *nfsd_serv;
	71
3fb803a9 AG	72	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	73	static struct svc_stat nfsd_acl_svcstats;
	74	static struct svc_version * nfsd_acl_version[] = {
	75	[2] = &nfsd_acl_version2,
	76	[3] = &nfsd_acl_version3,
	77	};
	78
	79	#define NFSD_ACL_MINVERS 2
e8c96f8c	80	#define NFSD_ACL_NRVERS ARRAY_SIZE(nfsd_acl_version)
3fb803a9 AG	81	static struct svc_version *nfsd_acl_versions[NFSD_ACL_NRVERS];
	82
	83	static struct svc_program nfsd_acl_program = {
	84	.pg_prog = NFS_ACL_PROGRAM,
	85	.pg_nvers = NFSD_ACL_NRVERS,
	86	.pg_vers = nfsd_acl_versions,
1a8eff6d	87	.pg_name = "nfsacl",
3fb803a9 AG	88	.pg_class = "nfsd",
	89	.pg_stats = &nfsd_acl_svcstats,
	90	.pg_authenticate = &svc_set_client,
	91	};
	92
	93	static struct svc_stat nfsd_acl_svcstats = {
	94	.program = &nfsd_acl_program,
	95	};
	96	#endif /* defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL) */
	97
70c3b76c N	98	static struct svc_version * nfsd_version[] = {
	99	[2] = &nfsd_version2,
	100	#if defined(CONFIG_NFSD_V3)
	101	[3] = &nfsd_version3,
	102	#endif
	103	#if defined(CONFIG_NFSD_V4)
	104	[4] = &nfsd_version4,
	105	#endif
	106	};
	107
	108	#define NFSD_MINVERS 2
e8c96f8c	109	#define NFSD_NRVERS ARRAY_SIZE(nfsd_version)
70c3b76c N	110	static struct svc_version *nfsd_versions[NFSD_NRVERS];
	111
	112	struct svc_program nfsd_program = {
3fb803a9 AG	113	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	114	.pg_next = &nfsd_acl_program,
	115	#endif
70c3b76c N	116	.pg_prog = NFS_PROGRAM, /* program number */
	117	.pg_nvers = NFSD_NRVERS, /* nr of entries in nfsd_version */
	118	.pg_vers = nfsd_versions, /* version table */
	119	.pg_name = "nfsd", /* program name */
	120	.pg_class = "nfsd", /* authentication class */
	121	.pg_stats = &nfsd_svcstats, /* version table */
	122	.pg_authenticate = &svc_set_client, /* export authentication */
	123
	124	};
	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	}
1da177e4 LT	152	/*
	153	* Maximum number of nfsd processes
	154	*/
	155	#define NFSD_MAXSERVS 8192
	156
	157	int nfsd_nrthreads(void)
	158	{
c7d106c9 NB	159	int rv = 0;
	160	mutex_lock(&nfsd_mutex);
	161	if (nfsd_serv)
	162	rv = nfsd_serv->sv_nrthreads;
	163	mutex_unlock(&nfsd_mutex);
	164	return rv;
1da177e4 LT	165	}
1da177e4 LT	166
bc591ccf N	167	static void nfsd_last_thread(struct svc_serv *serv)
	168	{
	169	/* When last nfsd thread exits we need to do some clean-up */
7a182083 TT	170	struct svc_xprt *xprt;
7a182083 TT	171	list_for_each_entry(xprt, &serv->sv_permsocks, xpt_list)
24e36663	172	lockd_down();
bc591ccf N	173	nfsd_serv = NULL;
	174	nfsd_racache_shutdown();
	175	nfs4_state_shutdown();
	176
e096bbc6 JL	177	printk(KERN_WARNING "nfsd: last server has exited, flushing export "
	178	"cache\n");
	179	nfsd_export_flush();
bc591ccf	180	}
6658d3a7 N	181
	182	void nfsd_reset_versions(void)
	183	{
	184	int found_one = 0;
	185	int i;
	186
	187	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++) {
	188	if (nfsd_program.pg_vers[i])
	189	found_one = 1;
	190	}
	191
	192	if (!found_one) {
	193	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++)
	194	nfsd_program.pg_vers[i] = nfsd_version[i];
	195	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	196	for (i = NFSD_ACL_MINVERS; i < NFSD_ACL_NRVERS; i++)
	197	nfsd_acl_program.pg_vers[i] =
	198	nfsd_acl_version[i];
	199	#endif
	200	}
	201	}
	202
bedbdd8b	203
b41b66d6	204	int nfsd_create_serv(void)
02a375f0 N	205	{
02a375f0 N	206	int err = 0;
bedbdd8b NB	207
bedbdd8b NB	208	WARN_ON(!mutex_is_locked(&nfsd_mutex));
02a375f0	209	if (nfsd_serv) {
9a24ab57	210	svc_get(nfsd_serv);
02a375f0 N	211	return 0;
02a375f0 N	212	}
596bbe53 N	213	if (nfsd_max_blksize == 0) {
	214	/* choose a suitable default */
	215	struct sysinfo i;
	216	si_meminfo(&i);
	217	/* Aim for 1/4096 of memory per thread
	218	* This gives 1MB on 4Gig machines
	219	* But only uses 32K on 128M machines.
	220	* Bottom out at 8K on 32M and smaller.
	221	* Of course, this is only a default.
	222	*/
	223	nfsd_max_blksize = NFSSVC_MAXBLKSIZE;
44c55600	224	i.totalram <<= PAGE_SHIFT - 12;
596bbe53 N	225	while (nfsd_max_blksize > i.totalram &&
	226	nfsd_max_blksize >= 810242)
	227	nfsd_max_blksize /= 2;
	228	}
02a375f0 N	229
02a375f0 N	230	atomic_set(&nfsd_busy, 0);
e096bbc6	231	nfsd_serv = svc_create_pooled(&nfsd_program, nfsd_max_blksize,
e851db5b	232	AF_INET,
a75c5d01	233	nfsd_last_thread, nfsd, THIS_MODULE);
02a375f0 N	234	if (nfsd_serv == NULL)
02a375f0 N	235	err = -ENOMEM;
bedbdd8b	236
02a375f0 N	237	do_gettimeofday(&nfssvc_boot); /* record boot time */
	238	return err;
	239	}
	240
	241	static int nfsd_init_socks(int port)
	242	{
	243	int error;
	244	if (!list_empty(&nfsd_serv->sv_permsocks))
	245	return 0;
	246
26a41409	247	error = svc_create_xprt(nfsd_serv, "udp", port,
482fb94e	248	SVC_SOCK_DEFAULTS);
02a375f0 N	249	if (error < 0)
	250	return error;
	251
1cd9cd16 CL	252	error = lockd_up();
	253	if (error < 0)
	254	return error;
	255
26a41409	256	error = svc_create_xprt(nfsd_serv, "tcp", port,
482fb94e	257	SVC_SOCK_DEFAULTS);
02a375f0 N	258	if (error < 0)
02a375f0 N	259	return error;
26a41409 CL	260
	261	error = lockd_up();
	262	if (error < 0)
	263	return error;
	264
02a375f0 N	265	return 0;
	266	}
	267
eed2965a GB	268	int nfsd_nrpools(void)
	269	{
	270	if (nfsd_serv == NULL)
	271	return 0;
	272	else
	273	return nfsd_serv->sv_nrpools;
	274	}
	275
	276	int nfsd_get_nrthreads(int n, int *nthreads)
	277	{
	278	int i = 0;
	279
	280	if (nfsd_serv != NULL) {
	281	for (i = 0; i < nfsd_serv->sv_nrpools && i < n; i++)
	282	nthreads[i] = nfsd_serv->sv_pools[i].sp_nrthreads;
	283	}
	284
	285	return 0;
	286	}
	287
	288	int nfsd_set_nrthreads(int n, int *nthreads)
	289	{
	290	int i = 0;
	291	int tot = 0;
	292	int err = 0;
	293
bedbdd8b NB	294	WARN_ON(!mutex_is_locked(&nfsd_mutex));
bedbdd8b NB	295
eed2965a GB	296	if (nfsd_serv == NULL \|\| n <= 0)
	297	return 0;
	298
	299	if (n > nfsd_serv->sv_nrpools)
	300	n = nfsd_serv->sv_nrpools;
	301
	302	/* enforce a global maximum number of threads */
	303	tot = 0;
	304	for (i = 0; i < n; i++) {
	305	if (nthreads[i] > NFSD_MAXSERVS)
	306	nthreads[i] = NFSD_MAXSERVS;
	307	tot += nthreads[i];
	308	}
	309	if (tot > NFSD_MAXSERVS) {
	310	/* total too large: scale down requested numbers */
	311	for (i = 0; i < n && tot > 0; i++) {
	312	int new = nthreads[i] * NFSD_MAXSERVS / tot;
	313	tot -= (nthreads[i] - new);
	314	nthreads[i] = new;
	315	}
	316	for (i = 0; i < n && tot > 0; i++) {
	317	nthreads[i]--;
	318	tot--;
	319	}
	320	}
	321
	322	/*
	323	* There must always be a thread in pool 0; the admin
	324	* can't shut down NFS completely using pool_threads.
	325	*/
	326	if (nthreads[0] == 0)
	327	nthreads[0] = 1;
	328
	329	/* apply the new numbers */
eed2965a GB	330	svc_get(nfsd_serv);
	331	for (i = 0; i < n; i++) {
	332	err = svc_set_num_threads(nfsd_serv, &nfsd_serv->sv_pools[i],
	333	nthreads[i]);
	334	if (err)
	335	break;
	336	}
	337	svc_destroy(nfsd_serv);
eed2965a GB	338
	339	return err;
	340	}
	341
1da177e4 LT	342	int
	343	nfsd_svc(unsigned short port, int nrservs)
	344	{
	345	int error;
bedbdd8b NB	346
bedbdd8b NB	347	mutex_lock(&nfsd_mutex);
6658d3a7	348	dprintk("nfsd: creating service\n");
1da177e4 LT	349	error = -EINVAL;
	350	if (nrservs <= 0)
	351	nrservs = 0;
	352	if (nrservs > NFSD_MAXSERVS)
	353	nrservs = NFSD_MAXSERVS;
	354
	355	/* Readahead param cache - will no-op if it already exists */
	356	error = nfsd_racache_init(2*nrservs);
	357	if (error<0)
	358	goto out;
e8ff2a84	359	nfs4_state_start();
02a375f0 N	360
	361	nfsd_reset_versions();
	362
	363	error = nfsd_create_serv();
	364
	365	if (error)
	366	goto out;
	367	error = nfsd_init_socks(port);
	368	if (error)
	369	goto failure;
	370
eec09661	371	error = svc_set_num_threads(nfsd_serv, NULL, nrservs);
1da177e4	372	failure:
1da177e4	373	svc_destroy(nfsd_serv); /* Release server */
1da177e4	374	out:
bedbdd8b	375	mutex_unlock(&nfsd_mutex);
1da177e4 LT	376	return error;
	377	}
	378
	379	static inline void
	380	update_thread_usage(int busy_threads)
	381	{
	382	unsigned long prev_call;
	383	unsigned long diff;
	384	int decile;
	385
	386	spin_lock(&nfsd_call_lock);
	387	prev_call = nfsd_last_call;
	388	nfsd_last_call = jiffies;
	389	decile = busy_threads*10/nfsdstats.th_cnt;
	390	if (decile>0 && decile <= 10) {
	391	diff = nfsd_last_call - prev_call;
	392	if ( (nfsdstats.th_usage[decile-1] += diff) >= NFSD_USAGE_WRAP)
	393	nfsdstats.th_usage[decile-1] -= NFSD_USAGE_WRAP;
	394	if (decile == 10)
	395	nfsdstats.th_fullcnt++;
	396	}
	397	spin_unlock(&nfsd_call_lock);
	398	}
	399
	400	/*
	401	* This is the NFS server kernel thread
	402	*/
9867d76c JL	403	static int
9867d76c JL	404	nfsd(void *vrqstp)
1da177e4	405	{
9867d76c	406	struct svc_rqst rqstp = (struct svc_rqst ) vrqstp;
9867d76c	407	int err, preverr = 0;
1da177e4 LT	408
1da177e4 LT	409	/* Lock module and set up kernel thread */
bedbdd8b	410	mutex_lock(&nfsd_mutex);
1da177e4	411
9867d76c	412	/* At this point, the thread shares current->fs
1da177e4 LT	413	* with the init process. We need to create files with a
1da177e4 LT	414	* umask of 0 instead of init's umask. */
3e93cd67	415	if (unshare_fs_struct() < 0) {
1da177e4 LT	416	printk("Unable to start nfsd thread: out of memory\n");
	417	goto out;
	418	}
3e93cd67	419
1da177e4 LT	420	current->fs->umask = 0;
1da177e4 LT	421
9867d76c JL	422	/*
9867d76c JL	423	* thread is spawned with all signals set to SIG_IGN, re-enable
100766f8	424	* the ones that will bring down the thread
9867d76c	425	*/
100766f8 JL	426	allow_signal(SIGKILL);
	427	allow_signal(SIGHUP);
	428	allow_signal(SIGINT);
	429	allow_signal(SIGQUIT);
bedbdd8b	430
1da177e4	431	nfsdstats.th_cnt++;
bedbdd8b NB	432	mutex_unlock(&nfsd_mutex);
bedbdd8b NB	433
1da177e4 LT	434	/*
	435	* We want less throttling in balance_dirty_pages() so that nfs to
	436	* localhost doesn't cause nfsd to lock up due to all the client's
	437	* dirty pages.
	438	*/
	439	current->flags \|= PF_LESS_THROTTLE;
83144186	440	set_freezable();
1da177e4 LT	441
	442	/*
	443	* The main request loop
	444	*/
	445	for (;;) {
1da177e4 LT	446	/*
	447	* Find a socket with data available and call its
	448	* recvfrom routine.
	449	*/
6fb2b47f	450	while ((err = svc_recv(rqstp, 6060HZ)) == -EAGAIN)
1da177e4	451	;
9867d76c	452	if (err == -EINTR)
1da177e4	453	break;
9867d76c JL	454	else if (err < 0) {
	455	if (err != preverr) {
	456	printk(KERN_WARNING "%s: unexpected error "
	457	"from svc_recv (%d)\n", __func__, -err);
	458	preverr = err;
	459	}
	460	schedule_timeout_uninterruptible(HZ);
	461	continue;
	462	}
	463
1da177e4 LT	464	update_thread_usage(atomic_read(&nfsd_busy));
	465	atomic_inc(&nfsd_busy);
	466
	467	/* Lock the export hash tables for reading. */
	468	exp_readlock();
	469
6fb2b47f	470	svc_process(rqstp);
1da177e4 LT	471
	472	/* Unlock export hash tables */
	473	exp_readunlock();
	474	update_thread_usage(atomic_read(&nfsd_busy));
	475	atomic_dec(&nfsd_busy);
	476	}
	477
24e36663	478	/* Clear signals before calling svc_exit_thread() */
9e416052	479	flush_signals(current);
1da177e4	480
bedbdd8b	481	mutex_lock(&nfsd_mutex);
1da177e4 LT	482	nfsdstats.th_cnt --;
	483
	484	out:
	485	/* Release the thread */
	486	svc_exit_thread(rqstp);
	487
	488	/* Release module */
bedbdd8b	489	mutex_unlock(&nfsd_mutex);
1da177e4	490	module_put_and_exit(0);
9867d76c	491	return 0;
1da177e4 LT	492	}
1da177e4 LT	493
32c1eb0c AA	494	static __be32 map_new_errors(u32 vers, __be32 nfserr)
	495	{
	496	if (nfserr == nfserr_jukebox && vers == 2)
	497	return nfserr_dropit;
	498	if (nfserr == nfserr_wrongsec && vers < 4)
	499	return nfserr_acces;
	500	return nfserr;
	501	}
	502
1da177e4	503	int
c7afef1f	504	nfsd_dispatch(struct svc_rqst rqstp, __be32 statp)
1da177e4 LT	505	{
	506	struct svc_procedure *proc;
	507	kxdrproc_t xdr;
ad451d38 AV	508	__be32 nfserr;
ad451d38 AV	509	__be32 *nfserrp;
1da177e4 LT	510
	511	dprintk("nfsd_dispatch: vers %d proc %d\n",
	512	rqstp->rq_vers, rqstp->rq_proc);
	513	proc = rqstp->rq_procinfo;
	514
	515	/* Check whether we have this call in the cache. */
	516	switch (nfsd_cache_lookup(rqstp, proc->pc_cachetype)) {
	517	case RC_INTR:
	518	case RC_DROPIT:
	519	return 0;
	520	case RC_REPLY:
	521	return 1;
	522	case RC_DOIT:;
	523	/* do it */
	524	}
	525
	526	/* Decode arguments */
	527	xdr = proc->pc_decode;
ad451d38	528	if (xdr && !xdr(rqstp, (__be32*)rqstp->rq_arg.head[0].iov_base,
1da177e4 LT	529	rqstp->rq_argp)) {
	530	dprintk("nfsd: failed to decode arguments!\n");
	531	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	532	*statp = rpc_garbage_args;
	533	return 1;
	534	}
	535
	536	/* need to grab the location to store the status, as
	537	* nfsv4 does some encoding while processing
	538	*/
	539	nfserrp = rqstp->rq_res.head[0].iov_base
	540	+ rqstp->rq_res.head[0].iov_len;
ad451d38	541	rqstp->rq_res.head[0].iov_len += sizeof(__be32);
1da177e4 LT	542
	543	/* Now call the procedure handler, and encode NFS status. */
	544	nfserr = proc->pc_func(rqstp, rqstp->rq_argp, rqstp->rq_resp);
32c1eb0c	545	nfserr = map_new_errors(rqstp->rq_vers, nfserr);
1da177e4	546	if (nfserr == nfserr_dropit) {
45457e09	547	dprintk("nfsd: Dropping request; may be revisited later\n");
1da177e4 LT	548	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	549	return 0;
	550	}
	551
	552	if (rqstp->rq_proc != 0)
	553	*nfserrp++ = nfserr;
	554
	555	/* Encode result.
	556	* For NFSv2, additional info is never returned in case of an error.
	557	*/
	558	if (!(nfserr && rqstp->rq_vers == 2)) {
	559	xdr = proc->pc_encode;
	560	if (xdr && !xdr(rqstp, nfserrp,
	561	rqstp->rq_resp)) {
	562	/* Failed to encode result. Release cache entry */
	563	dprintk("nfsd: failed to encode result!\n");
	564	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	565	*statp = rpc_system_err;
	566	return 1;
	567	}
	568	}
	569
	570	/* Store reply in cache. */
	571	nfsd_cache_update(rqstp, proc->pc_cachetype, statp + 1);
	572	return 1;
	573	}