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

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

#include <linux/sched.h>
#include <linux/freezer.h>
#include <linux/fs_struct.h>
#include <linux/swap.h>

#include <linux/sunrpc/stats.h>
#include <linux/sunrpc/svcsock.h>
#include <linux/lockd/bind.h>
#include <linux/nfsacl.h>
#include <linux/seq_file.h>
#include <net/net_namespace.h>
#include "nfsd.h"
#include "cache.h"
#include "vfs.h"

#define NFSDDBG_FACILITY	NFSDDBG_SVC

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

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

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

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

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

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

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

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

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

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

};

u32 nfsd_supported_minorversion;

int nfsd_vers(int vers, enum vers_op change)
{
	if (vers < NFSD_MINVERS || vers >= NFSD_NRVERS)
		return 0;
	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 int nfsd_init_socks(int port)
{
	int error;
	if (!list_empty(&nfsd_serv->sv_permsocks))
		return 0;

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

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

	return 0;
}

static bool nfsd_up = false;

static int nfsd_startup(unsigned short port, int nrservs)
{
	int ret;

	if (nfsd_up)
		return 0;
	/*
	 * Readahead param cache - will no-op if it already exists.
	 * (Note therefore results will be suboptimal if number of
	 * threads is modified after nfsd start.)
	 */
	ret = nfsd_racache_init(2*nrservs);
	if (ret)
		return ret;
	ret = nfsd_init_socks(port);
	if (ret)
		goto out_racache;
	ret = lockd_up();
	if (ret)
		goto out_racache;
	ret = nfs4_state_start();
	if (ret)
		goto out_lockd;
	nfsd_up = true;
	return 0;
out_lockd:
	lockd_down();
out_racache:
	nfsd_racache_shutdown();
	return ret;
}

static void nfsd_shutdown(void)
{
	/*
	 * write_ports can create the server without actually starting
	 * any threads--if we get shut down before any threads are
	 * started, then nfsd_last_thread will be run before any of this
	 * other initialization has been done.
	 */
	if (!nfsd_up)
		return;
	nfs4_state_shutdown();
	lockd_down();
	nfsd_racache_shutdown();
	nfsd_up = false;
}

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_shutdown();

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

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

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

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

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

int nfsd_create_serv(void)
{
	int err = 0;

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

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

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

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;
}

/*
 * Adjust the number of threads and return the new number of threads.
 * This is also the function that starts the server if necessary, if
 * this is the first time nrservs is nonzero.
 */
int
nfsd_svc(unsigned short port, int nrservs)
{
	int	error;
	bool	nfsd_up_before;

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

	error = nfsd_create_serv();
	if (error)
		goto out;

	nfsd_up_before = nfsd_up;

	error = nfsd_startup(port, nrservs);
	if (error)
		goto out_destroy;
	error = svc_set_num_threads(nfsd_serv, NULL, nrservs);
	if (error)
		goto out_shutdown;
	/* We are holding a reference to nfsd_serv which
	 * we don't want to count in the return value,
	 * so subtract 1
	 */
	error = nfsd_serv->sv_nrthreads - 1;
out_shutdown:
	if (error < 0 && !nfsd_up_before)
		nfsd_shutdown();
out_destroy:
	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();

		validate_process_creds();
		svc_process(rqstp);
		validate_process_creds();

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

int nfsd_pool_stats_release(struct inode *inode, struct file *file)
{
	int ret = seq_release(inode, file);
	mutex_lock(&nfsd_mutex);
	/* this function really, really should have been called svc_put() */
	svc_destroy(nfsd_serv);
	mutex_unlock(&nfsd_mutex);
	return ret;
}
Commit	Line	Data
1da177e4	1	/*
1da177e4 LT	2	* Central processing for nfsd.
	3	*
	4	* Authors: Olaf Kirch ([email protected])
	5	*
	6	* Copyright (C) 1995, 1996, 1997 Olaf Kirch <[email protected]>
	7	*/
	8
e8edc6e0	9	#include <linux/sched.h>
83144186	10	#include <linux/freezer.h>
1da177e4	11	#include <linux/fs_struct.h>
c3d06f9c	12	#include <linux/swap.h>
1da177e4	13
1da177e4	14	#include <linux/sunrpc/stats.h>
1da177e4	15	#include <linux/sunrpc/svcsock.h>
1da177e4	16	#include <linux/lockd/bind.h>
a257cdd0	17	#include <linux/nfsacl.h>
ed2d8aed	18	#include <linux/seq_file.h>
fc5d00b0	19	#include <net/net_namespace.h>
9a74af21 BH	20	#include "nfsd.h"
9a74af21 BH	21	#include "cache.h"
0a3adade	22	#include "vfs.h"
1da177e4 LT	23
	24	#define NFSDDBG_FACILITY NFSDDBG_SVC
	25
1da177e4	26	extern struct svc_program nfsd_program;
9867d76c	27	static int nfsd(void *vrqstp);
1da177e4	28	struct timeval nfssvc_boot;
1da177e4	29
bedbdd8b NB	30	/*
	31	* nfsd_mutex protects nfsd_serv -- both the pointer itself and the members
	32	* of the svc_serv struct. In particular, ->sv_nrthreads but also to some
	33	* extent ->sv_temp_socks and ->sv_permsocks. It also protects nfsdstats.th_cnt
	34	*
	35	* If (out side the lock) nfsd_serv is non-NULL, then it must point to a
	36	* properly initialised 'struct svc_serv' with ->sv_nrthreads > 0. That number
	37	* of nfsd threads must exist and each must listed in ->sp_all_threads in each
	38	* entry of ->sv_pools[].
	39	*
	40	* Transitions of the thread count between zero and non-zero are of particular
	41	* interest since the svc_serv needs to be created and initialized at that
	42	* point, or freed.
3dd98a3b JL	43	*
	44	* Finally, the nfsd_mutex also protects some of the global variables that are
	45	* accessed when nfsd starts and that are settable via the write_* routines in
	46	* nfsctl.c. In particular:
	47	*
	48	* user_recovery_dirname
	49	* user_lease_time
	50	* nfsd_versions
bedbdd8b NB	51	*/
	52	DEFINE_MUTEX(nfsd_mutex);
	53	struct svc_serv *nfsd_serv;
	54
4bd9b0f4 AA	55	/*
	56	* nfsd_drc_lock protects nfsd_drc_max_pages and nfsd_drc_pages_used.
	57	* nfsd_drc_max_pages limits the total amount of memory available for
	58	* version 4.1 DRC caches.
	59	* nfsd_drc_pages_used tracks the current version 4.1 DRC memory usage.
	60	*/
	61	spinlock_t nfsd_drc_lock;
0c193054 AA	62	unsigned int nfsd_drc_max_mem;
0c193054 AA	63	unsigned int nfsd_drc_mem_used;
4bd9b0f4	64
3fb803a9 AG	65	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	66	static struct svc_stat nfsd_acl_svcstats;
	67	static struct svc_version * nfsd_acl_version[] = {
	68	[2] = &nfsd_acl_version2,
	69	[3] = &nfsd_acl_version3,
	70	};
	71
	72	#define NFSD_ACL_MINVERS 2
e8c96f8c	73	#define NFSD_ACL_NRVERS ARRAY_SIZE(nfsd_acl_version)
3fb803a9 AG	74	static struct svc_version *nfsd_acl_versions[NFSD_ACL_NRVERS];
	75
	76	static struct svc_program nfsd_acl_program = {
	77	.pg_prog = NFS_ACL_PROGRAM,
	78	.pg_nvers = NFSD_ACL_NRVERS,
	79	.pg_vers = nfsd_acl_versions,
1a8eff6d	80	.pg_name = "nfsacl",
3fb803a9 AG	81	.pg_class = "nfsd",
	82	.pg_stats = &nfsd_acl_svcstats,
	83	.pg_authenticate = &svc_set_client,
	84	};
	85
	86	static struct svc_stat nfsd_acl_svcstats = {
	87	.program = &nfsd_acl_program,
	88	};
	89	#endif /* defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL) */
	90
70c3b76c N	91	static struct svc_version * nfsd_version[] = {
	92	[2] = &nfsd_version2,
	93	#if defined(CONFIG_NFSD_V3)
	94	[3] = &nfsd_version3,
	95	#endif
	96	#if defined(CONFIG_NFSD_V4)
	97	[4] = &nfsd_version4,
	98	#endif
	99	};
	100
	101	#define NFSD_MINVERS 2
e8c96f8c	102	#define NFSD_NRVERS ARRAY_SIZE(nfsd_version)
70c3b76c N	103	static struct svc_version *nfsd_versions[NFSD_NRVERS];
	104
	105	struct svc_program nfsd_program = {
3fb803a9 AG	106	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	107	.pg_next = &nfsd_acl_program,
	108	#endif
70c3b76c N	109	.pg_prog = NFS_PROGRAM, /* program number */
	110	.pg_nvers = NFSD_NRVERS, /* nr of entries in nfsd_version */
	111	.pg_vers = nfsd_versions, /* version table */
	112	.pg_name = "nfsd", /* program name */
	113	.pg_class = "nfsd", /* authentication class */
	114	.pg_stats = &nfsd_svcstats, /* version table */
	115	.pg_authenticate = &svc_set_client, /* export authentication */
	116
	117	};
	118
8daf220a BH	119	u32 nfsd_supported_minorversion;
8daf220a BH	120
6658d3a7 N	121	int nfsd_vers(int vers, enum vers_op change)
	122	{
	123	if (vers < NFSD_MINVERS \|\| vers >= NFSD_NRVERS)
15ddb4ae	124	return 0;
6658d3a7 N	125	switch(change) {
	126	case NFSD_SET:
	127	nfsd_versions[vers] = nfsd_version[vers];
6658d3a7 N	128	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
6658d3a7 N	129	if (vers < NFSD_ACL_NRVERS)
1a8eff6d	130	nfsd_acl_versions[vers] = nfsd_acl_version[vers];
6658d3a7	131	#endif
1a8eff6d	132	break;
6658d3a7 N	133	case NFSD_CLEAR:
	134	nfsd_versions[vers] = NULL;
	135	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	136	if (vers < NFSD_ACL_NRVERS)
1a8eff6d	137	nfsd_acl_versions[vers] = NULL;
6658d3a7 N	138	#endif
	139	break;
	140	case NFSD_TEST:
	141	return nfsd_versions[vers] != NULL;
	142	case NFSD_AVAIL:
	143	return nfsd_version[vers] != NULL;
	144	}
	145	return 0;
	146	}
8daf220a BH	147
	148	int nfsd_minorversion(u32 minorversion, enum vers_op change)
	149	{
	150	if (minorversion > NFSD_SUPPORTED_MINOR_VERSION)
	151	return -1;
	152	switch(change) {
	153	case NFSD_SET:
	154	nfsd_supported_minorversion = minorversion;
	155	break;
	156	case NFSD_CLEAR:
	157	if (minorversion == 0)
	158	return -1;
	159	nfsd_supported_minorversion = minorversion - 1;
	160	break;
	161	case NFSD_TEST:
	162	return minorversion <= nfsd_supported_minorversion;
	163	case NFSD_AVAIL:
	164	return minorversion <= NFSD_SUPPORTED_MINOR_VERSION;
	165	}
	166	return 0;
	167	}
	168
1da177e4 LT	169	/*
	170	* Maximum number of nfsd processes
	171	*/
	172	#define NFSD_MAXSERVS 8192
	173
	174	int nfsd_nrthreads(void)
	175	{
c7d106c9 NB	176	int rv = 0;
	177	mutex_lock(&nfsd_mutex);
	178	if (nfsd_serv)
	179	rv = nfsd_serv->sv_nrthreads;
	180	mutex_unlock(&nfsd_mutex);
	181	return rv;
1da177e4 LT	182	}
1da177e4 LT	183
59db4a0c BF	184	static int nfsd_init_socks(int port)
	185	{
	186	int error;
	187	if (!list_empty(&nfsd_serv->sv_permsocks))
	188	return 0;
	189
fc5d00b0	190	error = svc_create_xprt(nfsd_serv, "udp", &init_net, PF_INET, port,
59db4a0c BF	191	SVC_SOCK_DEFAULTS);
	192	if (error < 0)
	193	return error;
	194
fc5d00b0	195	error = svc_create_xprt(nfsd_serv, "tcp", &init_net, PF_INET, port,
59db4a0c BF	196	SVC_SOCK_DEFAULTS);
	197	if (error < 0)
	198	return error;
	199
	200	return 0;
	201	}
	202
4ad9a344 JL	203	static bool nfsd_up = false;
	204
	205	static int nfsd_startup(unsigned short port, int nrservs)
	206	{
	207	int ret;
774f8bbd BF	208
	209	if (nfsd_up)
	210	return 0;
59db4a0c BF	211	/*
	212	* Readahead param cache - will no-op if it already exists.
	213	* (Note therefore results will be suboptimal if number of
	214	* threads is modified after nfsd start.)
	215	*/
	216	ret = nfsd_racache_init(2*nrservs);
	217	if (ret)
	218	return ret;
	219	ret = nfsd_init_socks(port);
	220	if (ret)
	221	goto out_racache;
ac77efbe JL	222	ret = lockd_up();
ac77efbe JL	223	if (ret)
774f8bbd	224	goto out_racache;
4ad9a344	225	ret = nfs4_state_start();
ac77efbe JL	226	if (ret)
ac77efbe JL	227	goto out_lockd;
4ad9a344	228	nfsd_up = true;
ac77efbe JL	229	return 0;
	230	out_lockd:
	231	lockd_down();
59db4a0c BF	232	out_racache:
59db4a0c BF	233	nfsd_racache_shutdown();
4ad9a344 JL	234	return ret;
	235	}
	236
	237	static void nfsd_shutdown(void)
	238	{
	239	/*
	240	* write_ports can create the server without actually starting
	241	* any threads--if we get shut down before any threads are
	242	* started, then nfsd_last_thread will be run before any of this
	243	* other initialization has been done.
	244	*/
	245	if (!nfsd_up)
	246	return;
	247	nfs4_state_shutdown();
59db4a0c BF	248	lockd_down();
59db4a0c BF	249	nfsd_racache_shutdown();
4ad9a344 JL	250	nfsd_up = false;
	251	}
	252
bc591ccf N	253	static void nfsd_last_thread(struct svc_serv *serv)
	254	{
	255	/* When last nfsd thread exits we need to do some clean-up */
	256	nfsd_serv = NULL;
4ad9a344	257	nfsd_shutdown();
bc591ccf	258
e096bbc6 JL	259	printk(KERN_WARNING "nfsd: last server has exited, flushing export "
	260	"cache\n");
	261	nfsd_export_flush();
bc591ccf	262	}
6658d3a7 N	263
	264	void nfsd_reset_versions(void)
	265	{
	266	int found_one = 0;
	267	int i;
	268
	269	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++) {
	270	if (nfsd_program.pg_vers[i])
	271	found_one = 1;
	272	}
	273
	274	if (!found_one) {
	275	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++)
	276	nfsd_program.pg_vers[i] = nfsd_version[i];
	277	#if defined(CONFIG_NFSD_V2_ACL) \|\| defined(CONFIG_NFSD_V3_ACL)
	278	for (i = NFSD_ACL_MINVERS; i < NFSD_ACL_NRVERS; i++)
	279	nfsd_acl_program.pg_vers[i] =
	280	nfsd_acl_version[i];
	281	#endif
	282	}
	283	}
	284
c3d06f9c AA	285	/*
	286	* Each session guarantees a negotiated per slot memory cache for replies
	287	* which in turn consumes memory beyond the v2/v3/v4.0 server. A dedicated
	288	* NFSv4.1 server might want to use more memory for a DRC than a machine
	289	* with mutiple services.
	290	*
	291	* Impose a hard limit on the number of pages for the DRC which varies
	292	* according to the machines free pages. This is of course only a default.
	293	*
	294	* For now this is a #defined shift which could be under admin control
	295	* in the future.
	296	*/
	297	static void set_max_drc(void)
	298	{
6a14dd1a	299	#define NFSD_DRC_SIZE_SHIFT 10
0c193054 AA	300	nfsd_drc_max_mem = (nr_free_buffer_pages()
	301	>> NFSD_DRC_SIZE_SHIFT) * PAGE_SIZE;
	302	nfsd_drc_mem_used = 0;
4bd9b0f4	303	spin_lock_init(&nfsd_drc_lock);
0c193054	304	dprintk("%s nfsd_drc_max_mem %u \n", __func__, nfsd_drc_max_mem);
c3d06f9c	305	}
bedbdd8b	306
b41b66d6	307	int nfsd_create_serv(void)
02a375f0 N	308	{
02a375f0 N	309	int err = 0;
bedbdd8b NB	310
bedbdd8b NB	311	WARN_ON(!mutex_is_locked(&nfsd_mutex));
02a375f0	312	if (nfsd_serv) {
9a24ab57	313	svc_get(nfsd_serv);
02a375f0 N	314	return 0;
02a375f0 N	315	}
596bbe53 N	316	if (nfsd_max_blksize == 0) {
	317	/* choose a suitable default */
	318	struct sysinfo i;
	319	si_meminfo(&i);
	320	/* Aim for 1/4096 of memory per thread
	321	* This gives 1MB on 4Gig machines
	322	* But only uses 32K on 128M machines.
	323	* Bottom out at 8K on 32M and smaller.
	324	* Of course, this is only a default.
	325	*/
	326	nfsd_max_blksize = NFSSVC_MAXBLKSIZE;
44c55600	327	i.totalram <<= PAGE_SHIFT - 12;
596bbe53 N	328	while (nfsd_max_blksize > i.totalram &&
	329	nfsd_max_blksize >= 810242)
	330	nfsd_max_blksize /= 2;
	331	}
e844a7b9	332	nfsd_reset_versions();
02a375f0	333
e096bbc6	334	nfsd_serv = svc_create_pooled(&nfsd_program, nfsd_max_blksize,
a75c5d01	335	nfsd_last_thread, nfsd, THIS_MODULE);
02a375f0	336	if (nfsd_serv == NULL)
628b3687	337	return -ENOMEM;
bedbdd8b	338
628b3687	339	set_max_drc();
02a375f0 N	340	do_gettimeofday(&nfssvc_boot); /* record boot time */
	341	return err;
	342	}
	343
eed2965a GB	344	int nfsd_nrpools(void)
	345	{
	346	if (nfsd_serv == NULL)
	347	return 0;
	348	else
	349	return nfsd_serv->sv_nrpools;
	350	}
	351
	352	int nfsd_get_nrthreads(int n, int *nthreads)
	353	{
	354	int i = 0;
	355
	356	if (nfsd_serv != NULL) {
	357	for (i = 0; i < nfsd_serv->sv_nrpools && i < n; i++)
	358	nthreads[i] = nfsd_serv->sv_pools[i].sp_nrthreads;
	359	}
	360
	361	return 0;
	362	}
	363
	364	int nfsd_set_nrthreads(int n, int *nthreads)
	365	{
	366	int i = 0;
	367	int tot = 0;
	368	int err = 0;
	369
bedbdd8b NB	370	WARN_ON(!mutex_is_locked(&nfsd_mutex));
bedbdd8b NB	371
eed2965a GB	372	if (nfsd_serv == NULL \|\| n <= 0)
	373	return 0;
	374
	375	if (n > nfsd_serv->sv_nrpools)
	376	n = nfsd_serv->sv_nrpools;
	377
	378	/* enforce a global maximum number of threads */
	379	tot = 0;
	380	for (i = 0; i < n; i++) {
	381	if (nthreads[i] > NFSD_MAXSERVS)
	382	nthreads[i] = NFSD_MAXSERVS;
	383	tot += nthreads[i];
	384	}
	385	if (tot > NFSD_MAXSERVS) {
	386	/* total too large: scale down requested numbers */
	387	for (i = 0; i < n && tot > 0; i++) {
	388	int new = nthreads[i] * NFSD_MAXSERVS / tot;
	389	tot -= (nthreads[i] - new);
	390	nthreads[i] = new;
	391	}
	392	for (i = 0; i < n && tot > 0; i++) {
	393	nthreads[i]--;
	394	tot--;
	395	}
	396	}
	397
	398	/*
	399	* There must always be a thread in pool 0; the admin
	400	* can't shut down NFS completely using pool_threads.
	401	*/
	402	if (nthreads[0] == 0)
	403	nthreads[0] = 1;
	404
	405	/* apply the new numbers */
eed2965a GB	406	svc_get(nfsd_serv);
	407	for (i = 0; i < n; i++) {
	408	err = svc_set_num_threads(nfsd_serv, &nfsd_serv->sv_pools[i],
	409	nthreads[i]);
	410	if (err)
	411	break;
	412	}
	413	svc_destroy(nfsd_serv);
eed2965a GB	414
	415	return err;
	416	}
	417
ac77efbe JL	418	/*
	419	* Adjust the number of threads and return the new number of threads.
	420	* This is also the function that starts the server if necessary, if
	421	* this is the first time nrservs is nonzero.
	422	*/
1da177e4 LT	423	int
	424	nfsd_svc(unsigned short port, int nrservs)
	425	{
	426	int error;
774f8bbd	427	bool nfsd_up_before;
bedbdd8b NB	428
bedbdd8b NB	429	mutex_lock(&nfsd_mutex);
6658d3a7	430	dprintk("nfsd: creating service\n");
1da177e4 LT	431	if (nrservs <= 0)
	432	nrservs = 0;
	433	if (nrservs > NFSD_MAXSERVS)
	434	nrservs = NFSD_MAXSERVS;
671e1fcf N	435	error = 0;
	436	if (nrservs == 0 && nfsd_serv == NULL)
	437	goto out;
	438
02a375f0	439	error = nfsd_create_serv();
02a375f0	440	if (error)
774f8bbd BF	441	goto out;
	442
	443	nfsd_up_before = nfsd_up;
	444
	445	error = nfsd_startup(port, nrservs);
af4718f3 BF	446	if (error)
af4718f3 BF	447	goto out_destroy;
774f8bbd BF	448	error = svc_set_num_threads(nfsd_serv, NULL, nrservs);
	449	if (error)
	450	goto out_shutdown;
af4718f3 BF	451	/* We are holding a reference to nfsd_serv which
	452	* we don't want to count in the return value,
	453	* so subtract 1
	454	*/
	455	error = nfsd_serv->sv_nrthreads - 1;
4ad9a344	456	out_shutdown:
774f8bbd	457	if (error < 0 && !nfsd_up_before)
4ad9a344	458	nfsd_shutdown();
774f8bbd BF	459	out_destroy:
774f8bbd BF	460	svc_destroy(nfsd_serv); /* Release server */
4ad9a344	461	out:
bedbdd8b	462	mutex_unlock(&nfsd_mutex);
1da177e4 LT	463	return error;
	464	}
	465
1da177e4 LT	466
	467	/*
	468	* This is the NFS server kernel thread
	469	*/
9867d76c JL	470	static int
9867d76c JL	471	nfsd(void *vrqstp)
1da177e4	472	{
9867d76c	473	struct svc_rqst rqstp = (struct svc_rqst ) vrqstp;
9867d76c	474	int err, preverr = 0;
1da177e4 LT	475
1da177e4 LT	476	/* Lock module and set up kernel thread */
bedbdd8b	477	mutex_lock(&nfsd_mutex);
1da177e4	478
9867d76c	479	/* At this point, the thread shares current->fs
1da177e4 LT	480	* with the init process. We need to create files with a
1da177e4 LT	481	* umask of 0 instead of init's umask. */
3e93cd67	482	if (unshare_fs_struct() < 0) {
1da177e4 LT	483	printk("Unable to start nfsd thread: out of memory\n");
	484	goto out;
	485	}
3e93cd67	486
1da177e4 LT	487	current->fs->umask = 0;
1da177e4 LT	488
9867d76c JL	489	/*
9867d76c JL	490	* thread is spawned with all signals set to SIG_IGN, re-enable
100766f8	491	* the ones that will bring down the thread
9867d76c	492	*/
100766f8 JL	493	allow_signal(SIGKILL);
	494	allow_signal(SIGHUP);
	495	allow_signal(SIGINT);
	496	allow_signal(SIGQUIT);
bedbdd8b	497
1da177e4	498	nfsdstats.th_cnt++;
bedbdd8b NB	499	mutex_unlock(&nfsd_mutex);
bedbdd8b NB	500
1da177e4 LT	501	/*
	502	* We want less throttling in balance_dirty_pages() so that nfs to
	503	* localhost doesn't cause nfsd to lock up due to all the client's
	504	* dirty pages.
	505	*/
	506	current->flags \|= PF_LESS_THROTTLE;
83144186	507	set_freezable();
1da177e4 LT	508
	509	/*
	510	* The main request loop
	511	*/
	512	for (;;) {
1da177e4 LT	513	/*
	514	* Find a socket with data available and call its
	515	* recvfrom routine.
	516	*/
6fb2b47f	517	while ((err = svc_recv(rqstp, 6060HZ)) == -EAGAIN)
1da177e4	518	;
9867d76c	519	if (err == -EINTR)
1da177e4	520	break;
9867d76c JL	521	else if (err < 0) {
	522	if (err != preverr) {
	523	printk(KERN_WARNING "%s: unexpected error "
	524	"from svc_recv (%d)\n", __func__, -err);
	525	preverr = err;
	526	}
	527	schedule_timeout_uninterruptible(HZ);
	528	continue;
	529	}
	530
1da177e4 LT	531
	532	/* Lock the export hash tables for reading. */
	533	exp_readlock();
	534
e0e81739	535	validate_process_creds();
6fb2b47f	536	svc_process(rqstp);
e0e81739	537	validate_process_creds();
1da177e4 LT	538
	539	/* Unlock export hash tables */
	540	exp_readunlock();
1da177e4 LT	541	}
1da177e4 LT	542
24e36663	543	/* Clear signals before calling svc_exit_thread() */
9e416052	544	flush_signals(current);
1da177e4	545
bedbdd8b	546	mutex_lock(&nfsd_mutex);
1da177e4 LT	547	nfsdstats.th_cnt --;
	548
	549	out:
	550	/* Release the thread */
	551	svc_exit_thread(rqstp);
	552
	553	/* Release module */
bedbdd8b	554	mutex_unlock(&nfsd_mutex);
1da177e4	555	module_put_and_exit(0);
9867d76c	556	return 0;
1da177e4 LT	557	}
1da177e4 LT	558
32c1eb0c AA	559	static __be32 map_new_errors(u32 vers, __be32 nfserr)
	560	{
	561	if (nfserr == nfserr_jukebox && vers == 2)
	562	return nfserr_dropit;
	563	if (nfserr == nfserr_wrongsec && vers < 4)
	564	return nfserr_acces;
	565	return nfserr;
	566	}
	567
1da177e4	568	int
c7afef1f	569	nfsd_dispatch(struct svc_rqst rqstp, __be32 statp)
1da177e4 LT	570	{
	571	struct svc_procedure *proc;
	572	kxdrproc_t xdr;
ad451d38 AV	573	__be32 nfserr;
ad451d38 AV	574	__be32 *nfserrp;
1da177e4 LT	575
	576	dprintk("nfsd_dispatch: vers %d proc %d\n",
	577	rqstp->rq_vers, rqstp->rq_proc);
	578	proc = rqstp->rq_procinfo;
	579
	580	/* Check whether we have this call in the cache. */
	581	switch (nfsd_cache_lookup(rqstp, proc->pc_cachetype)) {
	582	case RC_INTR:
	583	case RC_DROPIT:
	584	return 0;
	585	case RC_REPLY:
	586	return 1;
	587	case RC_DOIT:;
	588	/* do it */
	589	}
	590
	591	/* Decode arguments */
	592	xdr = proc->pc_decode;
ad451d38	593	if (xdr && !xdr(rqstp, (__be32*)rqstp->rq_arg.head[0].iov_base,
1da177e4 LT	594	rqstp->rq_argp)) {
	595	dprintk("nfsd: failed to decode arguments!\n");
	596	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	597	*statp = rpc_garbage_args;
	598	return 1;
	599	}
	600
	601	/* need to grab the location to store the status, as
	602	* nfsv4 does some encoding while processing
	603	*/
	604	nfserrp = rqstp->rq_res.head[0].iov_base
	605	+ rqstp->rq_res.head[0].iov_len;
ad451d38	606	rqstp->rq_res.head[0].iov_len += sizeof(__be32);
1da177e4 LT	607
	608	/* Now call the procedure handler, and encode NFS status. */
	609	nfserr = proc->pc_func(rqstp, rqstp->rq_argp, rqstp->rq_resp);
32c1eb0c	610	nfserr = map_new_errors(rqstp->rq_vers, nfserr);
1da177e4	611	if (nfserr == nfserr_dropit) {
45457e09	612	dprintk("nfsd: Dropping request; may be revisited later\n");
1da177e4 LT	613	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	614	return 0;
	615	}
	616
	617	if (rqstp->rq_proc != 0)
	618	*nfserrp++ = nfserr;
	619
	620	/* Encode result.
	621	* For NFSv2, additional info is never returned in case of an error.
	622	*/
	623	if (!(nfserr && rqstp->rq_vers == 2)) {
	624	xdr = proc->pc_encode;
	625	if (xdr && !xdr(rqstp, nfserrp,
	626	rqstp->rq_resp)) {
	627	/* Failed to encode result. Release cache entry */
	628	dprintk("nfsd: failed to encode result!\n");
	629	nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
	630	*statp = rpc_system_err;
	631	return 1;
	632	}
	633	}
	634
	635	/* Store reply in cache. */
	636	nfsd_cache_update(rqstp, proc->pc_cachetype, statp + 1);
	637	return 1;
	638	}
03cf6c9f GB	639
	640	int nfsd_pool_stats_open(struct inode inode, struct file file)
	641	{
ed2d8aed RY	642	int ret;
	643	mutex_lock(&nfsd_mutex);
	644	if (nfsd_serv == NULL) {
	645	mutex_unlock(&nfsd_mutex);
03cf6c9f	646	return -ENODEV;
ed2d8aed RY	647	}
	648	/* bump up the psudo refcount while traversing */
	649	svc_get(nfsd_serv);
	650	ret = svc_pool_stats_open(nfsd_serv, file);
	651	mutex_unlock(&nfsd_mutex);
	652	return ret;
	653	}
	654
	655	int nfsd_pool_stats_release(struct inode inode, struct file file)
	656	{
	657	int ret = seq_release(inode, file);
	658	mutex_lock(&nfsd_mutex);
	659	/* this function really, really should have been called svc_put() */
	660	svc_destroy(nfsd_serv);
	661	mutex_unlock(&nfsd_mutex);
	662	return ret;
03cf6c9f	663	}