[GitHub/mt8127/android_kernel_alcatel_ttab.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/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);

#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)
{
	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 */
	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\n");
	if (killsig != SIG_NOCLEAN) {
		printk(KERN_WARNING "nfsd: unexporting all filesystems\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;
	lock_kernel();
	if (nfsd_serv) {
		svc_get(nfsd_serv);
		unlock_kernel();
		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,
				      nfsd_last_thread,
				      nfsd, SIG_NOCLEAN, THIS_MODULE);
	if (nfsd_serv == NULL)
		err = -ENOMEM;
	unlock_kernel();
	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 = lockd_up(IPPROTO_UDP);
	if (error >= 0) {
		error = svc_create_xprt(nfsd_serv, "udp", port,
					SVC_SOCK_DEFAULTS);
		if (error < 0)
			lockd_down();
	}
	if (error < 0)
		return error;

	error = lockd_up(IPPROTO_TCP);
	if (error >= 0) {
		error = svc_create_xprt(nfsd_serv, "tcp", port,
					SVC_SOCK_DEFAULTS);
		if (error < 0)
			lockd_down();
	}
	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;

	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 */
	lock_kernel();
	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);
	unlock_kernel();

	return err;
}

int
nfsd_svc(unsigned short port, int nrservs)
{
	int	error;
	
	lock_kernel();
	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:
	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 fs_struct *fsp;
	int		err;
	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++;

	rqstp->rq_task = current;

	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;
	set_freezable();

	/*
	 * 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(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(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 svc_exit_thread() */
	flush_signals(current);

	lock_kernel();

	nfsdstats.th_cnt --;

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

	/* Release module */
	unlock_kernel();
	module_put_and_exit(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 LT	23	#include <linux/fs_struct.h>
	24
	25	#include <linux/sunrpc/types.h>
	26	#include <linux/sunrpc/stats.h>
	27	#include <linux/sunrpc/svc.h>
	28	#include <linux/sunrpc/svcsock.h>
	29	#include <linux/sunrpc/cache.h>
	30	#include <linux/nfsd/nfsd.h>
	31	#include <linux/nfsd/stats.h>
	32	#include <linux/nfsd/cache.h>
70c3b76c	33	#include <linux/nfsd/syscall.h>
1da177e4	34	#include <linux/lockd/bind.h>
a257cdd0	35	#include <linux/nfsacl.h>
1da177e4 LT	36
	37	#define NFSDDBG_FACILITY NFSDDBG_SVC
	38
	39	/* these signals will be delivered to an nfsd thread
	40	* when handling a request
	41	*/
	42	#define ALLOWED_SIGS (sigmask(SIGKILL))
	43	/* these signals will be delivered to an nfsd thread
	44	* when not handling a request. i.e. when waiting
	45	*/
	46	#define SHUTDOWN_SIGS (sigmask(SIGKILL) \| sigmask(SIGHUP) \| sigmask(SIGINT) \| sigmask(SIGQUIT))
	47	/* if the last thread dies with SIGHUP, then the exports table is
	48	* left unchanged ( like 2.4-{0-9} ). Any other signal will clear
	49	* the exports table (like 2.2).
	50	*/
	51	#define SIG_NOCLEAN SIGHUP
	52
	53	extern struct svc_program nfsd_program;
	54	static void nfsd(struct svc_rqst *rqstp);
	55	struct timeval nfssvc_boot;
70c3b76c	56	struct svc_serv *nfsd_serv;
1da177e4 LT	57	static atomic_t nfsd_busy;
	58	static unsigned long nfsd_last_call;
	59	static DEFINE_SPINLOCK(nfsd_call_lock);
	60
3fb803a9 AG	61	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	62	static struct svc_stat nfsd_acl_svcstats;
	63	static struct svc_version * nfsd_acl_version[] = {
	64	[2] = &nfsd_acl_version2,
	65	[3] = &nfsd_acl_version3,
	66	};
	67
	68	#define NFSD_ACL_MINVERS 2
e8c96f8c	69	#define NFSD_ACL_NRVERS ARRAY_SIZE(nfsd_acl_version)
3fb803a9 AG	70	static struct svc_version *nfsd_acl_versions[NFSD_ACL_NRVERS];
	71
	72	static struct svc_program nfsd_acl_program = {
	73	.pg_prog = NFS_ACL_PROGRAM,
	74	.pg_nvers = NFSD_ACL_NRVERS,
	75	.pg_vers = nfsd_acl_versions,
1a8eff6d	76	.pg_name = "nfsacl",
3fb803a9 AG	77	.pg_class = "nfsd",
	78	.pg_stats = &nfsd_acl_svcstats,
	79	.pg_authenticate = &svc_set_client,
	80	};
	81
	82	static struct svc_stat nfsd_acl_svcstats = {
	83	.program = &nfsd_acl_program,
	84	};
	85	#endif /* defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL) */
	86
70c3b76c N	87	static struct svc_version * nfsd_version[] = {
	88	[2] = &nfsd_version2,
	89	#if defined(CONFIG_NFSD_V3)
	90	[3] = &nfsd_version3,
	91	#endif
	92	#if defined(CONFIG_NFSD_V4)
	93	[4] = &nfsd_version4,
	94	#endif
	95	};
	96
	97	#define NFSD_MINVERS 2
e8c96f8c	98	#define NFSD_NRVERS ARRAY_SIZE(nfsd_version)
70c3b76c N	99	static struct svc_version *nfsd_versions[NFSD_NRVERS];
	100
	101	struct svc_program nfsd_program = {
3fb803a9 AG	102	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	103	.pg_next = &nfsd_acl_program,
	104	#endif
70c3b76c N	105	.pg_prog = NFS_PROGRAM, /* program number */
	106	.pg_nvers = NFSD_NRVERS, /* nr of entries in nfsd_version */
	107	.pg_vers = nfsd_versions, /* version table */
	108	.pg_name = "nfsd", /* program name */
	109	.pg_class = "nfsd", /* authentication class */
	110	.pg_stats = &nfsd_svcstats, /* version table */
	111	.pg_authenticate = &svc_set_client, /* export authentication */
	112
	113	};
	114
6658d3a7 N	115	int nfsd_vers(int vers, enum vers_op change)
	116	{
	117	if (vers < NFSD_MINVERS \|\| vers >= NFSD_NRVERS)
	118	return -1;
	119	switch(change) {
	120	case NFSD_SET:
	121	nfsd_versions[vers] = nfsd_version[vers];
6658d3a7 N	122	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
6658d3a7 N	123	if (vers < NFSD_ACL_NRVERS)
1a8eff6d	124	nfsd_acl_versions[vers] = nfsd_acl_version[vers];
6658d3a7	125	#endif
1a8eff6d	126	break;
6658d3a7 N	127	case NFSD_CLEAR:
	128	nfsd_versions[vers] = NULL;
	129	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	130	if (vers < NFSD_ACL_NRVERS)
1a8eff6d	131	nfsd_acl_versions[vers] = NULL;
6658d3a7 N	132	#endif
	133	break;
	134	case NFSD_TEST:
	135	return nfsd_versions[vers] != NULL;
	136	case NFSD_AVAIL:
	137	return nfsd_version[vers] != NULL;
	138	}
	139	return 0;
	140	}
1da177e4 LT	141	/*
	142	* Maximum number of nfsd processes
	143	*/
	144	#define NFSD_MAXSERVS 8192
	145
	146	int nfsd_nrthreads(void)
	147	{
	148	if (nfsd_serv == NULL)
	149	return 0;
	150	else
	151	return nfsd_serv->sv_nrthreads;
	152	}
	153
bc591ccf N	154	static int killsig; /* signal that was used to kill last nfsd */
	155	static void nfsd_last_thread(struct svc_serv *serv)
	156	{
	157	/* When last nfsd thread exits we need to do some clean-up */
7a182083 TT	158	struct svc_xprt *xprt;
7a182083 TT	159	list_for_each_entry(xprt, &serv->sv_permsocks, xpt_list)
24e36663	160	lockd_down();
bc591ccf N	161	nfsd_serv = NULL;
	162	nfsd_racache_shutdown();
	163	nfs4_state_shutdown();
	164
	165	printk(KERN_WARNING "nfsd: last server has exited\n");
	166	if (killsig != SIG_NOCLEAN) {
	167	printk(KERN_WARNING "nfsd: unexporting all filesystems\n");
	168	nfsd_export_flush();
	169	}
	170	}
6658d3a7 N	171
	172	void nfsd_reset_versions(void)
	173	{
	174	int found_one = 0;
	175	int i;
	176
	177	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++) {
	178	if (nfsd_program.pg_vers[i])
	179	found_one = 1;
	180	}
	181
	182	if (!found_one) {
	183	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++)
	184	nfsd_program.pg_vers[i] = nfsd_version[i];
	185	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	186	for (i = NFSD_ACL_MINVERS; i < NFSD_ACL_NRVERS; i++)
	187	nfsd_acl_program.pg_vers[i] =
	188	nfsd_acl_version[i];
	189	#endif
	190	}
	191	}
	192
b41b66d6	193	int nfsd_create_serv(void)
02a375f0 N	194	{
	195	int err = 0;
	196	lock_kernel();
	197	if (nfsd_serv) {
9a24ab57	198	svc_get(nfsd_serv);
02a375f0 N	199	unlock_kernel();
	200	return 0;
	201	}
596bbe53 N	202	if (nfsd_max_blksize == 0) {
	203	/* choose a suitable default */
	204	struct sysinfo i;
	205	si_meminfo(&i);
	206	/* Aim for 1/4096 of memory per thread
	207	* This gives 1MB on 4Gig machines
	208	* But only uses 32K on 128M machines.
	209	* Bottom out at 8K on 32M and smaller.
	210	* Of course, this is only a default.
	211	*/
	212	nfsd_max_blksize = NFSSVC_MAXBLKSIZE;
44c55600	213	i.totalram <<= PAGE_SHIFT - 12;
596bbe53 N	214	while (nfsd_max_blksize > i.totalram &&
	215	nfsd_max_blksize >= 810242)
	216	nfsd_max_blksize /= 2;
	217	}
02a375f0 N	218
02a375f0 N	219	atomic_set(&nfsd_busy, 0);
596bbe53	220	nfsd_serv = svc_create_pooled(&nfsd_program,
c6b0a9f8	221	nfsd_max_blksize,
eec09661 GB	222	nfsd_last_thread,
eec09661 GB	223	nfsd, SIG_NOCLEAN, THIS_MODULE);
02a375f0 N	224	if (nfsd_serv == NULL)
02a375f0 N	225	err = -ENOMEM;
02a375f0 N	226	unlock_kernel();
	227	do_gettimeofday(&nfssvc_boot); /* record boot time */
	228	return err;
	229	}
	230
	231	static int nfsd_init_socks(int port)
	232	{
	233	int error;
	234	if (!list_empty(&nfsd_serv->sv_permsocks))
	235	return 0;
	236
02a375f0	237	error = lockd_up(IPPROTO_UDP);
4a3ae42d	238	if (error >= 0) {
d7c9f1ed	239	error = svc_create_xprt(nfsd_serv, "udp", port,
482fb94e	240	SVC_SOCK_DEFAULTS);
4a3ae42d N	241	if (error < 0)
	242	lockd_down();
	243	}
02a375f0 N	244	if (error < 0)
	245	return error;
	246
02a375f0	247	error = lockd_up(IPPROTO_TCP);
4a3ae42d	248	if (error >= 0) {
d7c9f1ed	249	error = svc_create_xprt(nfsd_serv, "tcp", port,
482fb94e	250	SVC_SOCK_DEFAULTS);
4a3ae42d N	251	if (error < 0)
	252	lockd_down();
	253	}
02a375f0 N	254	if (error < 0)
02a375f0 N	255	return error;
02a375f0 N	256	return 0;
	257	}
	258
eed2965a GB	259	int nfsd_nrpools(void)
	260	{
	261	if (nfsd_serv == NULL)
	262	return 0;
	263	else
	264	return nfsd_serv->sv_nrpools;
	265	}
	266
	267	int nfsd_get_nrthreads(int n, int *nthreads)
	268	{
	269	int i = 0;
	270
	271	if (nfsd_serv != NULL) {
	272	for (i = 0; i < nfsd_serv->sv_nrpools && i < n; i++)
	273	nthreads[i] = nfsd_serv->sv_pools[i].sp_nrthreads;
	274	}
	275
	276	return 0;
	277	}
	278
	279	int nfsd_set_nrthreads(int n, int *nthreads)
	280	{
	281	int i = 0;
	282	int tot = 0;
	283	int err = 0;
	284
	285	if (nfsd_serv == NULL \|\| n <= 0)
	286	return 0;
	287
	288	if (n > nfsd_serv->sv_nrpools)
	289	n = nfsd_serv->sv_nrpools;
	290
	291	/* enforce a global maximum number of threads */
	292	tot = 0;
	293	for (i = 0; i < n; i++) {
	294	if (nthreads[i] > NFSD_MAXSERVS)
	295	nthreads[i] = NFSD_MAXSERVS;
	296	tot += nthreads[i];
	297	}
	298	if (tot > NFSD_MAXSERVS) {
	299	/* total too large: scale down requested numbers */
	300	for (i = 0; i < n && tot > 0; i++) {
	301	int new = nthreads[i] * NFSD_MAXSERVS / tot;
	302	tot -= (nthreads[i] - new);
	303	nthreads[i] = new;
	304	}
	305	for (i = 0; i < n && tot > 0; i++) {
	306	nthreads[i]--;
	307	tot--;
	308	}
	309	}
	310
	311	/*
	312	* There must always be a thread in pool 0; the admin
	313	* can't shut down NFS completely using pool_threads.
	314	*/
	315	if (nthreads[0] == 0)
	316	nthreads[0] = 1;
	317
	318	/* apply the new numbers */
	319	lock_kernel();
	320	svc_get(nfsd_serv);
	321	for (i = 0; i < n; i++) {
	322	err = svc_set_num_threads(nfsd_serv, &nfsd_serv->sv_pools[i],
323	nthreads[i]);
324	if (err)
325	break;
326	}
327	svc_destroy(nfsd_serv);
328	unlock_kernel();
329
330	return err;
331	}
332
1da177e4 LT	333	int
	334	nfsd_svc(unsigned short port, int nrservs)
	335	{
	336	int error;
1da177e4 LT	337
1da177e4 LT	338	lock_kernel();
6658d3a7	339	dprintk("nfsd: creating service\n");
1da177e4 LT	340	error = -EINVAL;
	341	if (nrservs <= 0)
	342	nrservs = 0;
	343	if (nrservs > NFSD_MAXSERVS)
	344	nrservs = NFSD_MAXSERVS;
	345
	346	/* Readahead param cache - will no-op if it already exists */
	347	error = nfsd_racache_init(2*nrservs);
	348	if (error<0)
	349	goto out;
e8ff2a84	350	nfs4_state_start();
02a375f0 N	351
	352	nfsd_reset_versions();
	353
	354	error = nfsd_create_serv();
	355
	356	if (error)
	357	goto out;
	358	error = nfsd_init_socks(port);
	359	if (error)
	360	goto failure;
	361
eec09661	362	error = svc_set_num_threads(nfsd_serv, NULL, nrservs);
1da177e4	363	failure:
1da177e4	364	svc_destroy(nfsd_serv); /* Release server */
1da177e4 LT	365	out:
	366	unlock_kernel();
	367	return error;
	368	}
	369
	370	static inline void
	371	update_thread_usage(int busy_threads)
	372	{
	373	unsigned long prev_call;
	374	unsigned long diff;
	375	int decile;
	376
	377	spin_lock(&nfsd_call_lock);
	378	prev_call = nfsd_last_call;
	379	nfsd_last_call = jiffies;
	380	decile = busy_threads*10/nfsdstats.th_cnt;
	381	if (decile>0 && decile <= 10) {
	382	diff = nfsd_last_call - prev_call;
	383	if ( (nfsdstats.th_usage[decile-1] += diff) >= NFSD_USAGE_WRAP)
	384	nfsdstats.th_usage[decile-1] -= NFSD_USAGE_WRAP;
	385	if (decile == 10)
	386	nfsdstats.th_fullcnt++;
	387	}
	388	spin_unlock(&nfsd_call_lock);
	389	}
	390
	391	/*
	392	* This is the NFS server kernel thread
	393	*/
	394	static void
	395	nfsd(struct svc_rqst *rqstp)
	396	{
1da177e4 LT	397	struct fs_struct *fsp;
1da177e4 LT	398	int err;
1da177e4 LT	399	sigset_t shutdown_mask, allowed_mask;
	400
	401	/* Lock module and set up kernel thread */
	402	lock_kernel();
	403	daemonize("nfsd");
	404
	405	/* After daemonize() this kernel thread shares current->fs
	406	* with the init process. We need to create files with a
	407	* umask of 0 instead of init's umask. */
	408	fsp = copy_fs_struct(current->fs);
	409	if (!fsp) {
	410	printk("Unable to start nfsd thread: out of memory\n");
	411	goto out;
	412	}
	413	exit_fs(current);
	414	current->fs = fsp;
	415	current->fs->umask = 0;
	416
	417	siginitsetinv(&shutdown_mask, SHUTDOWN_SIGS);
	418	siginitsetinv(&allowed_mask, ALLOWED_SIGS);
	419
	420	nfsdstats.th_cnt++;
	421
eec09661	422	rqstp->rq_task = current;
1da177e4 LT	423
	424	unlock_kernel();
	425
	426	/*
	427	* We want less throttling in balance_dirty_pages() so that nfs to
	428	* localhost doesn't cause nfsd to lock up due to all the client's
	429	* dirty pages.
	430	*/
	431	current->flags \|= PF_LESS_THROTTLE;
83144186	432	set_freezable();
1da177e4 LT	433
	434	/*
	435	* The main request loop
	436	*/
	437	for (;;) {
	438	/* Block all but the shutdown signals */
	439	sigprocmask(SIG_SETMASK, &shutdown_mask, NULL);
	440
	441	/*
	442	* Find a socket with data available and call its
	443	* recvfrom routine.
	444	*/
6fb2b47f	445	while ((err = svc_recv(rqstp, 6060HZ)) == -EAGAIN)
1da177e4 LT	446	;
	447	if (err < 0)
	448	break;
	449	update_thread_usage(atomic_read(&nfsd_busy));
	450	atomic_inc(&nfsd_busy);
	451
	452	/* Lock the export hash tables for reading. */
	453	exp_readlock();
	454
	455	/* Process request with signals blocked. */
	456	sigprocmask(SIG_SETMASK, &allowed_mask, NULL);
	457
6fb2b47f	458	svc_process(rqstp);
1da177e4 LT	459
	460	/* Unlock export hash tables */
	461	exp_readunlock();
	462	update_thread_usage(atomic_read(&nfsd_busy));
	463	atomic_dec(&nfsd_busy);
	464	}
	465
	466	if (err != -EINTR) {
	467	printk(KERN_WARNING "nfsd: terminating on error %d\n", -err);
	468	} else {
	469	unsigned int signo;
	470
	471	for (signo = 1; signo <= _NSIG; signo++)
	472	if (sigismember(&current->pending.signal, signo) &&
	473	!sigismember(&current->blocked, signo))
	474	break;
bc591ccf	475	killsig = signo;
1da177e4	476	}
24e36663	477	/* Clear signals before calling svc_exit_thread() */
9e416052	478	flush_signals(current);
1da177e4 LT	479
	480	lock_kernel();
	481
1da177e4 LT	482	nfsdstats.th_cnt --;
	483
	484	out:
	485	/* Release the thread */
	486	svc_exit_thread(rqstp);
	487
	488	/* Release module */
c4f92dba	489	unlock_kernel();
1da177e4 LT	490	module_put_and_exit(0);
	491	}
	492
32c1eb0c AA	493	static __be32 map_new_errors(u32 vers, __be32 nfserr)
	494	{
	495	if (nfserr == nfserr_jukebox && vers == 2)
	496	return nfserr_dropit;
	497	if (nfserr == nfserr_wrongsec && vers < 4)
	498	return nfserr_acces;
	499	return nfserr;
	500	}
	501
1da177e4	502	int
c7afef1f	503	nfsd_dispatch(struct svc_rqst rqstp, __be32 statp)
1da177e4 LT	504	{
	505	struct svc_procedure *proc;
	506	kxdrproc_t xdr;
ad451d38 AV	507	__be32 nfserr;
ad451d38 AV	508	__be32 *nfserrp;
1da177e4 LT	509
	510	dprintk("nfsd_dispatch: vers %d proc %d\n",
	511	rqstp->rq_vers, rqstp->rq_proc);
	512	proc = rqstp->rq_procinfo;
	513
	514	/* Check whether we have this call in the cache. */
	515	switch (nfsd_cache_lookup(rqstp, proc->pc_cachetype)) {
	516	case RC_INTR:
	517	case RC_DROPIT:
	518	return 0;
	519	case RC_REPLY:
	520	return 1;
	521	case RC_DOIT:;
	522	/* do it */
	523	}
	524
	525	/* Decode arguments */
	526	xdr = proc->pc_decode;
ad451d38	527	if (xdr && !xdr(rqstp, (__be32*)rqstp->rq_arg.head[0].iov_base,
1da177e4 LT	528	rqstp->rq_argp)) {
	529	dprintk("nfsd: failed to decode arguments!\n");
	530	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	531	*statp = rpc_garbage_args;
	532	return 1;
	533	}
	534
	535	/* need to grab the location to store the status, as
	536	* nfsv4 does some encoding while processing
	537	*/
	538	nfserrp = rqstp->rq_res.head[0].iov_base
	539	+ rqstp->rq_res.head[0].iov_len;
ad451d38	540	rqstp->rq_res.head[0].iov_len += sizeof(__be32);
1da177e4 LT	541
	542	/* Now call the procedure handler, and encode NFS status. */
	543	nfserr = proc->pc_func(rqstp, rqstp->rq_argp, rqstp->rq_resp);
32c1eb0c	544	nfserr = map_new_errors(rqstp->rq_vers, nfserr);
1da177e4	545	if (nfserr == nfserr_dropit) {
45457e09	546	dprintk("nfsd: Dropping request; may be revisited later\n");
1da177e4 LT	547	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	548	return 0;
	549	}
	550
	551	if (rqstp->rq_proc != 0)
	552	*nfserrp++ = nfserr;
	553
	554	/* Encode result.
	555	* For NFSv2, additional info is never returned in case of an error.
	556	*/
	557	if (!(nfserr && rqstp->rq_vers == 2)) {
	558	xdr = proc->pc_encode;
	559	if (xdr && !xdr(rqstp, nfserrp,
	560	rqstp->rq_resp)) {
	561	/* Failed to encode result. Release cache entry */
	562	dprintk("nfsd: failed to encode result!\n");
	563	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	564	*statp = rpc_system_err;
	565	return 1;
	566	}
	567	}
	568
	569	/* Store reply in cache. */
	570	nfsd_cache_update(rqstp, proc->pc_cachetype, statp + 1);
	571	return 1;
	572	}