[deliverable/linux.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/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/fs_struct.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

/* these signals will be delivered to an nfsd thread 
 * when handling a request
 */
#define ALLOWED_SIGS	(sigmask(SIGKILL))
/* these signals will be delivered to an nfsd thread
 * when not handling a request. i.e. when waiting
 */
#define SHUTDOWN_SIGS	(sigmask(SIGKILL) | sigmask(SIGHUP) | sigmask(SIGINT) | sigmask(SIGQUIT))
/* if the last thread dies with SIGHUP, then the exports table is
 * left unchanged ( like 2.4-{0-9} ).  Any other signal will clear
 * the exports table (like 2.2).
 */
#define	SIG_NOCLEAN	SIGHUP

extern struct svc_program	nfsd_program;
static void			nfsd(struct svc_rqst *rqstp);
struct timeval			nfssvc_boot;
       struct svc_serv 		*nfsd_serv;
static atomic_t			nfsd_busy;
static unsigned long		nfsd_last_call;
static DEFINE_SPINLOCK(nfsd_call_lock);

struct nfsd_list {
	struct list_head 	list;
	struct task_struct	*task;
};
static struct list_head nfsd_list = LIST_HEAD_INIT(nfsd_list);

#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		= "nfsd",
	.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 */

};

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

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

static int killsig;	/* signal that was used to kill last nfsd */
static void nfsd_last_thread(struct svc_serv *serv)
{
	/* When last nfsd thread exits we need to do some clean-up */
	nfsd_serv = NULL;
	nfsd_racache_shutdown();
	nfs4_state_shutdown();

	printk(KERN_WARNING "nfsd: last server has exited\n");
	if (killsig != SIG_NOCLEAN) {
		printk(KERN_WARNING "nfsd: unexporting all filesystems\n");
		nfsd_export_flush();
	}
}
int
nfsd_svc(unsigned short port, int nrservs)
{
	int	error;
	int	found_one, i;
	struct list_head *victim;
	
	lock_kernel();
	dprintk("nfsd: creating service: vers 0x%x\n",
		nfsd_versbits);
	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;
	error = nfs4_state_start();
	if (error<0)
		goto out;
	if (!nfsd_serv) {
		/*
		 * Use the nfsd_ctlbits to define which
		 * versions that will be advertised.
		 * If nfsd_ctlbits doesn't list any version,
		 * export them all.
		 */
		found_one = 0;

		for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++) {
			if (NFSCTL_VERISSET(nfsd_versbits, i)) {
				nfsd_program.pg_vers[i] = nfsd_version[i];
				found_one = 1;
			} else
				nfsd_program.pg_vers[i] = NULL;
		}

		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)
		found_one = 0;

		for (i = NFSD_ACL_MINVERS; i < NFSD_ACL_NRVERS; i++) {
			if (NFSCTL_VERISSET(nfsd_versbits, i)) {
				nfsd_acl_program.pg_vers[i] =
					nfsd_acl_version[i];
				found_one = 1;
			} else
				nfsd_acl_program.pg_vers[i] = NULL;
		}

		if (!found_one) {
			for (i = NFSD_ACL_MINVERS; i < NFSD_ACL_NRVERS; i++)
				nfsd_acl_program.pg_vers[i] =
					nfsd_acl_version[i];
		}
#endif

		atomic_set(&nfsd_busy, 0);
		error = -ENOMEM;
		nfsd_serv = svc_create(&nfsd_program, NFSD_BUFSIZE,
				       nfsd_last_thread);
		if (nfsd_serv == NULL)
			goto out;
		error = svc_makesock(nfsd_serv, IPPROTO_UDP, port);
		if (error < 0)
			goto failure;

#ifdef CONFIG_NFSD_TCP
		error = svc_makesock(nfsd_serv, IPPROTO_TCP, port);
		if (error < 0)
			goto failure;
#endif
		do_gettimeofday(&nfssvc_boot);		/* record boot time */
	} else
		nfsd_serv->sv_nrthreads++;
	nrservs -= (nfsd_serv->sv_nrthreads-1);
	while (nrservs > 0) {
		nrservs--;
		__module_get(THIS_MODULE);
		error = svc_create_thread(nfsd, nfsd_serv);
		if (error < 0) {
			module_put(THIS_MODULE);
			break;
		}
	}
	victim = nfsd_list.next;
	while (nrservs < 0 && victim != &nfsd_list) {
		struct nfsd_list *nl =
			list_entry(victim,struct nfsd_list, list);
		victim = victim->next;
		send_sig(SIG_NOCLEAN, nl->task, 1);
		nrservs++;
	}
 failure:
	svc_destroy(nfsd_serv);		/* Release server */
 out:
	unlock_kernel();
	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 void
nfsd(struct svc_rqst *rqstp)
{
	struct svc_serv	*serv = rqstp->rq_server;
	struct fs_struct *fsp;
	int		err;
	struct nfsd_list me;
	sigset_t shutdown_mask, allowed_mask;

	/* Lock module and set up kernel thread */
	lock_kernel();
	daemonize("nfsd");

	/* After daemonize() this kernel thread shares current->fs
	 * with the init process. We need to create files with a
	 * umask of 0 instead of init's umask. */
	fsp = copy_fs_struct(current->fs);
	if (!fsp) {
		printk("Unable to start nfsd thread: out of memory\n");
		goto out;
	}
	exit_fs(current);
	current->fs = fsp;
	current->fs->umask = 0;

	siginitsetinv(&shutdown_mask, SHUTDOWN_SIGS);
	siginitsetinv(&allowed_mask, ALLOWED_SIGS);

	nfsdstats.th_cnt++;

	lockd_up();				/* start lockd */

	me.task = current;
	list_add(&me.list, &nfsd_list);

	unlock_kernel();

	/*
	 * 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;

	/*
	 * The main request loop
	 */
	for (;;) {
		/* Block all but the shutdown signals */
		sigprocmask(SIG_SETMASK, &shutdown_mask, NULL);

		/*
		 * Find a socket with data available and call its
		 * recvfrom routine.
		 */
		while ((err = svc_recv(serv, rqstp,
				       60*60*HZ)) == -EAGAIN)
			;
		if (err < 0)
			break;
		update_thread_usage(atomic_read(&nfsd_busy));
		atomic_inc(&nfsd_busy);

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

		/* Process request with signals blocked.  */
		sigprocmask(SIG_SETMASK, &allowed_mask, NULL);

		svc_process(serv, rqstp);

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

	if (err != -EINTR) {
		printk(KERN_WARNING "nfsd: terminating on error %d\n", -err);
	} else {
		unsigned int	signo;

		for (signo = 1; signo <= _NSIG; signo++)
			if (sigismember(&current->pending.signal, signo) &&
			    !sigismember(&current->blocked, signo))
				break;
		killsig = signo;
	}
	/* Clear signals before calling lockd_down() and svc_exit_thread() */
	flush_signals(current);

	lock_kernel();

	/* Release lockd */
	lockd_down();
	list_del(&me.list);
	nfsdstats.th_cnt --;

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

	/* Release module */
	unlock_kernel();
	module_put_and_exit(0);
}

int
nfsd_dispatch(struct svc_rqst *rqstp, u32 *statp)
{
	struct svc_procedure	*proc;
	kxdrproc_t		xdr;
	u32			nfserr;
	u32			*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, (u32*)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(u32);

	/* Now call the procedure handler, and encode NFS status. */
	nfserr = proc->pc_func(rqstp, rqstp->rq_argp, rqstp->rq_resp);
	if (nfserr == nfserr_jukebox && rqstp->rq_vers == 2)
		nfserr = nfserr_dropit;
	if (nfserr == nfserr_dropit) {
		dprintk("nfsd: Dropping request due to malloc failure!\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 LT	11	#include <linux/module.h>
	12
	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>
	22	#include <linux/fs_struct.h>
	23
	24	#include <linux/sunrpc/types.h>
	25	#include <linux/sunrpc/stats.h>
	26	#include <linux/sunrpc/svc.h>
	27	#include <linux/sunrpc/svcsock.h>
	28	#include <linux/sunrpc/cache.h>
	29	#include <linux/nfsd/nfsd.h>
	30	#include <linux/nfsd/stats.h>
	31	#include <linux/nfsd/cache.h>
70c3b76c	32	#include <linux/nfsd/syscall.h>
1da177e4	33	#include <linux/lockd/bind.h>
a257cdd0	34	#include <linux/nfsacl.h>
1da177e4 LT	35
	36	#define NFSDDBG_FACILITY NFSDDBG_SVC
	37
	38	/* these signals will be delivered to an nfsd thread
	39	* when handling a request
	40	*/
	41	#define ALLOWED_SIGS (sigmask(SIGKILL))
	42	/* these signals will be delivered to an nfsd thread
	43	* when not handling a request. i.e. when waiting
	44	*/
	45	#define SHUTDOWN_SIGS (sigmask(SIGKILL) \| sigmask(SIGHUP) \| sigmask(SIGINT) \| sigmask(SIGQUIT))
	46	/* if the last thread dies with SIGHUP, then the exports table is
	47	* left unchanged ( like 2.4-{0-9} ). Any other signal will clear
	48	* the exports table (like 2.2).
	49	*/
	50	#define SIG_NOCLEAN SIGHUP
	51
	52	extern struct svc_program nfsd_program;
	53	static void nfsd(struct svc_rqst *rqstp);
	54	struct timeval nfssvc_boot;
70c3b76c	55	struct svc_serv *nfsd_serv;
1da177e4 LT	56	static atomic_t nfsd_busy;
	57	static unsigned long nfsd_last_call;
	58	static DEFINE_SPINLOCK(nfsd_call_lock);
	59
	60	struct nfsd_list {
	61	struct list_head list;
	62	struct task_struct *task;
	63	};
	64	static struct list_head nfsd_list = LIST_HEAD_INIT(nfsd_list);
	65
3fb803a9 AG	66	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	67	static struct svc_stat nfsd_acl_svcstats;
	68	static struct svc_version * nfsd_acl_version[] = {
	69	[2] = &nfsd_acl_version2,
	70	[3] = &nfsd_acl_version3,
	71	};
	72
	73	#define NFSD_ACL_MINVERS 2
e8c96f8c	74	#define NFSD_ACL_NRVERS ARRAY_SIZE(nfsd_acl_version)
3fb803a9 AG	75	static struct svc_version *nfsd_acl_versions[NFSD_ACL_NRVERS];
	76
	77	static struct svc_program nfsd_acl_program = {
	78	.pg_prog = NFS_ACL_PROGRAM,
	79	.pg_nvers = NFSD_ACL_NRVERS,
	80	.pg_vers = nfsd_acl_versions,
	81	.pg_name = "nfsd",
	82	.pg_class = "nfsd",
	83	.pg_stats = &nfsd_acl_svcstats,
	84	.pg_authenticate = &svc_set_client,
	85	};
	86
	87	static struct svc_stat nfsd_acl_svcstats = {
	88	.program = &nfsd_acl_program,
	89	};
	90	#endif /* defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL) */
	91
70c3b76c N	92	static struct svc_version * nfsd_version[] = {
	93	[2] = &nfsd_version2,
	94	#if defined(CONFIG_NFSD_V3)
	95	[3] = &nfsd_version3,
	96	#endif
	97	#if defined(CONFIG_NFSD_V4)
	98	[4] = &nfsd_version4,
	99	#endif
	100	};
	101
	102	#define NFSD_MINVERS 2
e8c96f8c	103	#define NFSD_NRVERS ARRAY_SIZE(nfsd_version)
70c3b76c N	104	static struct svc_version *nfsd_versions[NFSD_NRVERS];
	105
	106	struct svc_program nfsd_program = {
3fb803a9 AG	107	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	108	.pg_next = &nfsd_acl_program,
	109	#endif
70c3b76c N	110	.pg_prog = NFS_PROGRAM, /* program number */
	111	.pg_nvers = NFSD_NRVERS, /* nr of entries in nfsd_version */
	112	.pg_vers = nfsd_versions, /* version table */
	113	.pg_name = "nfsd", /* program name */
	114	.pg_class = "nfsd", /* authentication class */
	115	.pg_stats = &nfsd_svcstats, /* version table */
	116	.pg_authenticate = &svc_set_client, /* export authentication */
	117
	118	};
	119
1da177e4 LT	120	/*
	121	* Maximum number of nfsd processes
	122	*/
	123	#define NFSD_MAXSERVS 8192
	124
	125	int nfsd_nrthreads(void)
	126	{
	127	if (nfsd_serv == NULL)
	128	return 0;
	129	else
	130	return nfsd_serv->sv_nrthreads;
	131	}
	132
bc591ccf N	133	static int killsig; /* signal that was used to kill last nfsd */
	134	static void nfsd_last_thread(struct svc_serv *serv)
	135	{
	136	/* When last nfsd thread exits we need to do some clean-up */
	137	nfsd_serv = NULL;
	138	nfsd_racache_shutdown();
	139	nfs4_state_shutdown();
	140
	141	printk(KERN_WARNING "nfsd: last server has exited\n");
	142	if (killsig != SIG_NOCLEAN) {
	143	printk(KERN_WARNING "nfsd: unexporting all filesystems\n");
	144	nfsd_export_flush();
	145	}
	146	}
1da177e4 LT	147	int
	148	nfsd_svc(unsigned short port, int nrservs)
	149	{
	150	int error;
bc591ccf	151	int found_one, i;
1da177e4 LT	152	struct list_head *victim;
	153
	154	lock_kernel();
70c3b76c N	155	dprintk("nfsd: creating service: vers 0x%x\n",
70c3b76c N	156	nfsd_versbits);
1da177e4 LT	157	error = -EINVAL;
	158	if (nrservs <= 0)
	159	nrservs = 0;
	160	if (nrservs > NFSD_MAXSERVS)
	161	nrservs = NFSD_MAXSERVS;
	162
	163	/* Readahead param cache - will no-op if it already exists */
	164	error = nfsd_racache_init(2*nrservs);
	165	if (error<0)
	166	goto out;
76a3550e	167	error = nfs4_state_start();
1da177e4 LT	168	if (error<0)
	169	goto out;
	170	if (!nfsd_serv) {
70c3b76c N	171	/*
	172	* Use the nfsd_ctlbits to define which
	173	* versions that will be advertised.
	174	* If nfsd_ctlbits doesn't list any version,
	175	* export them all.
	176	*/
	177	found_one = 0;
	178
	179	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++) {
	180	if (NFSCTL_VERISSET(nfsd_versbits, i)) {
	181	nfsd_program.pg_vers[i] = nfsd_version[i];
	182	found_one = 1;
	183	} else
	184	nfsd_program.pg_vers[i] = NULL;
	185	}
	186
	187	if (!found_one) {
	188	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++)
	189	nfsd_program.pg_vers[i] = nfsd_version[i];
	190	}
	191
3fb803a9 AG	192
	193	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	194	found_one = 0;
	195
	196	for (i = NFSD_ACL_MINVERS; i < NFSD_ACL_NRVERS; i++) {
	197	if (NFSCTL_VERISSET(nfsd_versbits, i)) {
	198	nfsd_acl_program.pg_vers[i] =
	199	nfsd_acl_version[i];
	200	found_one = 1;
	201	} else
	202	nfsd_acl_program.pg_vers[i] = NULL;
	203	}
	204
	205	if (!found_one) {
	206	for (i = NFSD_ACL_MINVERS; i < NFSD_ACL_NRVERS; i++)
	207	nfsd_acl_program.pg_vers[i] =
	208	nfsd_acl_version[i];
	209	}
	210	#endif
	211
1da177e4 LT	212	atomic_set(&nfsd_busy, 0);
1da177e4 LT	213	error = -ENOMEM;
bc591ccf N	214	nfsd_serv = svc_create(&nfsd_program, NFSD_BUFSIZE,
bc591ccf N	215	nfsd_last_thread);
1da177e4 LT	216	if (nfsd_serv == NULL)
	217	goto out;
	218	error = svc_makesock(nfsd_serv, IPPROTO_UDP, port);
	219	if (error < 0)
	220	goto failure;
	221
	222	#ifdef CONFIG_NFSD_TCP
	223	error = svc_makesock(nfsd_serv, IPPROTO_TCP, port);
	224	if (error < 0)
	225	goto failure;
	226	#endif
	227	do_gettimeofday(&nfssvc_boot); /* record boot time */
	228	} else
	229	nfsd_serv->sv_nrthreads++;
	230	nrservs -= (nfsd_serv->sv_nrthreads-1);
	231	while (nrservs > 0) {
	232	nrservs--;
	233	__module_get(THIS_MODULE);
	234	error = svc_create_thread(nfsd, nfsd_serv);
	235	if (error < 0) {
	236	module_put(THIS_MODULE);
	237	break;
	238	}
	239	}
	240	victim = nfsd_list.next;
	241	while (nrservs < 0 && victim != &nfsd_list) {
	242	struct nfsd_list *nl =
	243	list_entry(victim,struct nfsd_list, list);
	244	victim = victim->next;
	245	send_sig(SIG_NOCLEAN, nl->task, 1);
	246	nrservs++;
	247	}
	248	failure:
1da177e4	249	svc_destroy(nfsd_serv); /* Release server */
1da177e4 LT	250	out:
	251	unlock_kernel();
	252	return error;
	253	}
	254
	255	static inline void
	256	update_thread_usage(int busy_threads)
	257	{
	258	unsigned long prev_call;
	259	unsigned long diff;
	260	int decile;
	261
	262	spin_lock(&nfsd_call_lock);
	263	prev_call = nfsd_last_call;
	264	nfsd_last_call = jiffies;
	265	decile = busy_threads*10/nfsdstats.th_cnt;
	266	if (decile>0 && decile <= 10) {
	267	diff = nfsd_last_call - prev_call;
	268	if ( (nfsdstats.th_usage[decile-1] += diff) >= NFSD_USAGE_WRAP)
	269	nfsdstats.th_usage[decile-1] -= NFSD_USAGE_WRAP;
	270	if (decile == 10)
	271	nfsdstats.th_fullcnt++;
	272	}
	273	spin_unlock(&nfsd_call_lock);
	274	}
	275
	276	/*
	277	* This is the NFS server kernel thread
	278	*/
	279	static void
	280	nfsd(struct svc_rqst *rqstp)
	281	{
	282	struct svc_serv *serv = rqstp->rq_server;
	283	struct fs_struct *fsp;
	284	int err;
	285	struct nfsd_list me;
	286	sigset_t shutdown_mask, allowed_mask;
	287
	288	/* Lock module and set up kernel thread */
	289	lock_kernel();
	290	daemonize("nfsd");
	291
	292	/* After daemonize() this kernel thread shares current->fs
	293	* with the init process. We need to create files with a
	294	* umask of 0 instead of init's umask. */
	295	fsp = copy_fs_struct(current->fs);
	296	if (!fsp) {
	297	printk("Unable to start nfsd thread: out of memory\n");
	298	goto out;
	299	}
	300	exit_fs(current);
	301	current->fs = fsp;
	302	current->fs->umask = 0;
	303
	304	siginitsetinv(&shutdown_mask, SHUTDOWN_SIGS);
	305	siginitsetinv(&allowed_mask, ALLOWED_SIGS);
	306
	307	nfsdstats.th_cnt++;
	308
	309	lockd_up(); /* start lockd */
	310
	311	me.task = current;
	312	list_add(&me.list, &nfsd_list);
	313
314	unlock_kernel();
315
316	/*
317	* We want less throttling in balance_dirty_pages() so that nfs to
318	* localhost doesn't cause nfsd to lock up due to all the client's
319	* dirty pages.
320	*/
321	current->flags \|= PF_LESS_THROTTLE;
322
323	/*
324	* The main request loop
325	*/
326	for (;;) {
327	/* Block all but the shutdown signals */
328	sigprocmask(SIG_SETMASK, &shutdown_mask, NULL);
329
330	/*
331	* Find a socket with data available and call its
332	* recvfrom routine.
333	*/
334	while ((err = svc_recv(serv, rqstp,
335	6060HZ)) == -EAGAIN)
336	;
337	if (err < 0)
338	break;
339	update_thread_usage(atomic_read(&nfsd_busy));
340	atomic_inc(&nfsd_busy);
341
342	/* Lock the export hash tables for reading. */
343	exp_readlock();
344
345	/* Process request with signals blocked. */
346	sigprocmask(SIG_SETMASK, &allowed_mask, NULL);
347
348	svc_process(serv, rqstp);
349
350	/* Unlock export hash tables */
351	exp_readunlock();
352	update_thread_usage(atomic_read(&nfsd_busy));
353	atomic_dec(&nfsd_busy);
354	}
355
356	if (err != -EINTR) {
357	printk(KERN_WARNING "nfsd: terminating on error %d\n", -err);
358	} else {
359	unsigned int signo;
360
361	for (signo = 1; signo <= _NSIG; signo++)
362	if (sigismember(&current->pending.signal, signo) &&
363	!sigismember(&current->blocked, signo))
364	break;
bc591ccf	365	killsig = signo;
1da177e4	366	}
9e416052 N	367	/* Clear signals before calling lockd_down() and svc_exit_thread() */
9e416052 N	368	flush_signals(current);
1da177e4 LT	369
	370	lock_kernel();
	371
	372	/* Release lockd */
	373	lockd_down();
1da177e4 LT	374	list_del(&me.list);
	375	nfsdstats.th_cnt --;
	376
	377	out:
	378	/* Release the thread */
	379	svc_exit_thread(rqstp);
	380
	381	/* Release module */
c4f92dba	382	unlock_kernel();
1da177e4 LT	383	module_put_and_exit(0);
	384	}
	385
	386	int
	387	nfsd_dispatch(struct svc_rqst rqstp, u32 statp)
	388	{
	389	struct svc_procedure *proc;
	390	kxdrproc_t xdr;
	391	u32 nfserr;
	392	u32 *nfserrp;
	393
	394	dprintk("nfsd_dispatch: vers %d proc %d\n",
	395	rqstp->rq_vers, rqstp->rq_proc);
	396	proc = rqstp->rq_procinfo;
	397
	398	/* Check whether we have this call in the cache. */
	399	switch (nfsd_cache_lookup(rqstp, proc->pc_cachetype)) {
	400	case RC_INTR:
	401	case RC_DROPIT:
	402	return 0;
	403	case RC_REPLY:
	404	return 1;
	405	case RC_DOIT:;
	406	/* do it */
	407	}
	408
	409	/* Decode arguments */
	410	xdr = proc->pc_decode;
	411	if (xdr && !xdr(rqstp, (u32*)rqstp->rq_arg.head[0].iov_base,
	412	rqstp->rq_argp)) {
	413	dprintk("nfsd: failed to decode arguments!\n");
	414	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	415	*statp = rpc_garbage_args;
	416	return 1;
	417	}
	418
	419	/* need to grab the location to store the status, as
	420	* nfsv4 does some encoding while processing
	421	*/
	422	nfserrp = rqstp->rq_res.head[0].iov_base
	423	+ rqstp->rq_res.head[0].iov_len;
	424	rqstp->rq_res.head[0].iov_len += sizeof(u32);
	425
	426	/* Now call the procedure handler, and encode NFS status. */
	427	nfserr = proc->pc_func(rqstp, rqstp->rq_argp, rqstp->rq_resp);
	428	if (nfserr == nfserr_jukebox && rqstp->rq_vers == 2)
	429	nfserr = nfserr_dropit;
	430	if (nfserr == nfserr_dropit) {
	431	dprintk("nfsd: Dropping request due to malloc failure!\n");
	432	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	433	return 0;
	434	}
	435
	436	if (rqstp->rq_proc != 0)
	437	*nfserrp++ = nfserr;
	438
	439	/* Encode result.
	440	* For NFSv2, additional info is never returned in case of an error.
	441	*/
	442	if (!(nfserr && rqstp->rq_vers == 2)) {
	443	xdr = proc->pc_encode;
	444	if (xdr && !xdr(rqstp, nfserrp,
	445	rqstp->rq_resp)) {
	446	/* Failed to encode result. Release cache entry */
447	dprintk("nfsd: failed to encode result!\n");
448	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
449	*statp = rpc_system_err;
450	return 1;
451	}
452	}
453
454	/* Store reply in cache. */
455	nfsd_cache_update(rqstp, proc->pc_cachetype, statp + 1);
456	return 1;
457	}