[net-next-2.6.git] / fs / dlm / plock.c

/*
 * Copyright (C) 2005-2008 Red Hat, Inc.  All rights reserved.
 *
 * This copyrighted material is made available to anyone wishing to use,
 * modify, copy, or redistribute it subject to the terms and conditions
 * of the GNU General Public License version 2.
 */

#include <linux/fs.h>
#include <linux/miscdevice.h>
#include <linux/poll.h>
#include <linux/dlm.h>
#include <linux/dlm_plock.h>
#include <linux/slab.h>

#include "dlm_internal.h"
#include "lockspace.h"

static spinlock_t ops_lock;
static struct list_head send_list;
static struct list_head recv_list;
static wait_queue_head_t send_wq;
static wait_queue_head_t recv_wq;

struct plock_op {
	struct list_head list;
	int done;
	struct dlm_plock_info info;
};

struct plock_xop {
	struct plock_op xop;
	void *callback;
	void *fl;
	void *file;
	struct file_lock flc;
};


static inline void set_version(struct dlm_plock_info *info)
{
	info->version[0] = DLM_PLOCK_VERSION_MAJOR;
	info->version[1] = DLM_PLOCK_VERSION_MINOR;
	info->version[2] = DLM_PLOCK_VERSION_PATCH;
}

static int check_version(struct dlm_plock_info *info)
{
	if ((DLM_PLOCK_VERSION_MAJOR != info->version[0]) ||
	    (DLM_PLOCK_VERSION_MINOR < info->version[1])) {
		log_print("plock device version mismatch: "
			  "kernel (%u.%u.%u), user (%u.%u.%u)",
			  DLM_PLOCK_VERSION_MAJOR,
			  DLM_PLOCK_VERSION_MINOR,
			  DLM_PLOCK_VERSION_PATCH,
			  info->version[0],
			  info->version[1],
			  info->version[2]);
		return -EINVAL;
	}
	return 0;
}

static void send_op(struct plock_op *op)
{
	set_version(&op->info);
	INIT_LIST_HEAD(&op->list);
	spin_lock(&ops_lock);
	list_add_tail(&op->list, &send_list);
	spin_unlock(&ops_lock);
	wake_up(&send_wq);
}

int dlm_posix_lock(dlm_lockspace_t *lockspace, u64 number, struct file *file,
		   int cmd, struct file_lock *fl)
{
	struct dlm_ls *ls;
	struct plock_op *op;
	struct plock_xop *xop;
	int rv;

	ls = dlm_find_lockspace_local(lockspace);
	if (!ls)
		return -EINVAL;

	xop = kzalloc(sizeof(*xop), GFP_NOFS);
	if (!xop) {
		rv = -ENOMEM;
		goto out;
	}

	op = &xop->xop;
	op->info.optype		= DLM_PLOCK_OP_LOCK;
	op->info.pid		= fl->fl_pid;
	op->info.ex		= (fl->fl_type == F_WRLCK);
	op->info.wait		= IS_SETLKW(cmd);
	op->info.fsid		= ls->ls_global_id;
	op->info.number		= number;
	op->info.start		= fl->fl_start;
	op->info.end		= fl->fl_end;
	if (fl->fl_lmops && fl->fl_lmops->fl_grant) {
		/* fl_owner is lockd which doesn't distinguish
		   processes on the nfs client */
		op->info.owner	= (__u64) fl->fl_pid;
		xop->callback	= fl->fl_lmops->fl_grant;
		locks_init_lock(&xop->flc);
		locks_copy_lock(&xop->flc, fl);
		xop->fl		= fl;
		xop->file	= file;
	} else {
		op->info.owner	= (__u64)(long) fl->fl_owner;
		xop->callback	= NULL;
	}

	send_op(op);

	if (xop->callback == NULL)
		wait_event(recv_wq, (op->done != 0));
	else {
		rv = FILE_LOCK_DEFERRED;
		goto out;
	}

	spin_lock(&ops_lock);
	if (!list_empty(&op->list)) {
		log_error(ls, "dlm_posix_lock: op on list %llx",
			  (unsigned long long)number);
		list_del(&op->list);
	}
	spin_unlock(&ops_lock);

	rv = op->info.rv;

	if (!rv) {
		if (posix_lock_file_wait(file, fl) < 0)
			log_error(ls, "dlm_posix_lock: vfs lock error %llx",
				  (unsigned long long)number);
	}

	kfree(xop);
out:
	dlm_put_lockspace(ls);
	return rv;
}
EXPORT_SYMBOL_GPL(dlm_posix_lock);

/* Returns failure iff a successful lock operation should be canceled */
static int dlm_plock_callback(struct plock_op *op)
{
	struct file *file;
	struct file_lock *fl;
	struct file_lock *flc;
	int (*notify)(void *, void *, int) = NULL;
	struct plock_xop *xop = (struct plock_xop *)op;
	int rv = 0;

	spin_lock(&ops_lock);
	if (!list_empty(&op->list)) {
		log_print("dlm_plock_callback: op on list %llx",
			  (unsigned long long)op->info.number);
		list_del(&op->list);
	}
	spin_unlock(&ops_lock);

	/* check if the following 2 are still valid or make a copy */
	file = xop->file;
	flc = &xop->flc;
	fl = xop->fl;
	notify = xop->callback;

	if (op->info.rv) {
		notify(fl, NULL, op->info.rv);
		goto out;
	}

	/* got fs lock; bookkeep locally as well: */
	flc->fl_flags &= ~FL_SLEEP;
	if (posix_lock_file(file, flc, NULL)) {
		/*
		 * This can only happen in the case of kmalloc() failure.
		 * The filesystem's own lock is the authoritative lock,
		 * so a failure to get the lock locally is not a disaster.
		 * As long as the fs cannot reliably cancel locks (especially
		 * in a low-memory situation), we're better off ignoring
		 * this failure than trying to recover.
		 */
		log_print("dlm_plock_callback: vfs lock error %llx file %p fl %p",
			  (unsigned long long)op->info.number, file, fl);
	}

	rv = notify(fl, NULL, 0);
	if (rv) {
		/* XXX: We need to cancel the fs lock here: */
		log_print("dlm_plock_callback: lock granted after lock request "
			  "failed; dangling lock!\n");
		goto out;
	}

out:
	kfree(xop);
	return rv;
}

int dlm_posix_unlock(dlm_lockspace_t *lockspace, u64 number, struct file *file,
		     struct file_lock *fl)
{
	struct dlm_ls *ls;
	struct plock_op *op;
	int rv;

	ls = dlm_find_lockspace_local(lockspace);
	if (!ls)
		return -EINVAL;

	op = kzalloc(sizeof(*op), GFP_NOFS);
	if (!op) {
		rv = -ENOMEM;
		goto out;
	}

	if (posix_lock_file_wait(file, fl) < 0)
		log_error(ls, "dlm_posix_unlock: vfs unlock error %llx",
			  (unsigned long long)number);

	op->info.optype		= DLM_PLOCK_OP_UNLOCK;
	op->info.pid		= fl->fl_pid;
	op->info.fsid		= ls->ls_global_id;
	op->info.number		= number;
	op->info.start		= fl->fl_start;
	op->info.end		= fl->fl_end;
	if (fl->fl_lmops && fl->fl_lmops->fl_grant)
		op->info.owner	= (__u64) fl->fl_pid;
	else
		op->info.owner	= (__u64)(long) fl->fl_owner;

	send_op(op);
	wait_event(recv_wq, (op->done != 0));

	spin_lock(&ops_lock);
	if (!list_empty(&op->list)) {
		log_error(ls, "dlm_posix_unlock: op on list %llx",
			  (unsigned long long)number);
		list_del(&op->list);
	}
	spin_unlock(&ops_lock);

	rv = op->info.rv;

	if (rv == -ENOENT)
		rv = 0;

	kfree(op);
out:
	dlm_put_lockspace(ls);
	return rv;
}
EXPORT_SYMBOL_GPL(dlm_posix_unlock);

int dlm_posix_get(dlm_lockspace_t *lockspace, u64 number, struct file *file,
		  struct file_lock *fl)
{
	struct dlm_ls *ls;
	struct plock_op *op;
	int rv;

	ls = dlm_find_lockspace_local(lockspace);
	if (!ls)
		return -EINVAL;

	op = kzalloc(sizeof(*op), GFP_NOFS);
	if (!op) {
		rv = -ENOMEM;
		goto out;
	}

	op->info.optype		= DLM_PLOCK_OP_GET;
	op->info.pid		= fl->fl_pid;
	op->info.ex		= (fl->fl_type == F_WRLCK);
	op->info.fsid		= ls->ls_global_id;
	op->info.number		= number;
	op->info.start		= fl->fl_start;
	op->info.end		= fl->fl_end;
	if (fl->fl_lmops && fl->fl_lmops->fl_grant)
		op->info.owner	= (__u64) fl->fl_pid;
	else
		op->info.owner	= (__u64)(long) fl->fl_owner;

	send_op(op);
	wait_event(recv_wq, (op->done != 0));

	spin_lock(&ops_lock);
	if (!list_empty(&op->list)) {
		log_error(ls, "dlm_posix_get: op on list %llx",
			  (unsigned long long)number);
		list_del(&op->list);
	}
	spin_unlock(&ops_lock);

	/* info.rv from userspace is 1 for conflict, 0 for no-conflict,
	   -ENOENT if there are no locks on the file */

	rv = op->info.rv;

	fl->fl_type = F_UNLCK;
	if (rv == -ENOENT)
		rv = 0;
	else if (rv > 0) {
		locks_init_lock(fl);
		fl->fl_type = (op->info.ex) ? F_WRLCK : F_RDLCK;
		fl->fl_flags = FL_POSIX;
		fl->fl_pid = op->info.pid;
		fl->fl_start = op->info.start;
		fl->fl_end = op->info.end;
		rv = 0;
	}

	kfree(op);
out:
	dlm_put_lockspace(ls);
	return rv;
}
EXPORT_SYMBOL_GPL(dlm_posix_get);

/* a read copies out one plock request from the send list */
static ssize_t dev_read(struct file *file, char __user *u, size_t count,
			loff_t *ppos)
{
	struct dlm_plock_info info;
	struct plock_op *op = NULL;

	if (count < sizeof(info))
		return -EINVAL;

	spin_lock(&ops_lock);
	if (!list_empty(&send_list)) {
		op = list_entry(send_list.next, struct plock_op, list);
		list_move(&op->list, &recv_list);
		memcpy(&info, &op->info, sizeof(info));
	}
	spin_unlock(&ops_lock);

	if (!op)
		return -EAGAIN;

	if (copy_to_user(u, &info, sizeof(info)))
		return -EFAULT;
	return sizeof(info);
}

/* a write copies in one plock result that should match a plock_op
   on the recv list */
static ssize_t dev_write(struct file *file, const char __user *u, size_t count,
			 loff_t *ppos)
{
	struct dlm_plock_info info;
	struct plock_op *op;
	int found = 0, do_callback = 0;

	if (count != sizeof(info))
		return -EINVAL;

	if (copy_from_user(&info, u, sizeof(info)))
		return -EFAULT;

	if (check_version(&info))
		return -EINVAL;

	spin_lock(&ops_lock);
	list_for_each_entry(op, &recv_list, list) {
		if (op->info.fsid == info.fsid &&
		    op->info.number == info.number &&
		    op->info.owner == info.owner) {
			struct plock_xop *xop = (struct plock_xop *)op;
			list_del_init(&op->list);
			memcpy(&op->info, &info, sizeof(info));
			if (xop->callback)
				do_callback = 1;
			else
				op->done = 1;
			found = 1;
			break;
		}
	}
	spin_unlock(&ops_lock);

	if (found) {
		if (do_callback)
			dlm_plock_callback(op);
		else
			wake_up(&recv_wq);
	} else
		log_print("dev_write no op %x %llx", info.fsid,
			  (unsigned long long)info.number);
	return count;
}

static unsigned int dev_poll(struct file *file, poll_table *wait)
{
	unsigned int mask = 0;

	poll_wait(file, &send_wq, wait);

	spin_lock(&ops_lock);
	if (!list_empty(&send_list))
		mask = POLLIN | POLLRDNORM;
	spin_unlock(&ops_lock);

	return mask;
}

static const struct file_operations dev_fops = {
	.read    = dev_read,
	.write   = dev_write,
	.poll    = dev_poll,
	.owner   = THIS_MODULE,
	.llseek  = noop_llseek,
};

static struct miscdevice plock_dev_misc = {
	.minor = MISC_DYNAMIC_MINOR,
	.name = DLM_PLOCK_MISC_NAME,
	.fops = &dev_fops
};

int dlm_plock_init(void)
{
	int rv;

	spin_lock_init(&ops_lock);
	INIT_LIST_HEAD(&send_list);
	INIT_LIST_HEAD(&recv_list);
	init_waitqueue_head(&send_wq);
	init_waitqueue_head(&recv_wq);

	rv = misc_register(&plock_dev_misc);
	if (rv)
		log_print("dlm_plock_init: misc_register failed %d", rv);
	return rv;
}

void dlm_plock_exit(void)
{
	if (misc_deregister(&plock_dev_misc) < 0)
		log_print("dlm_plock_exit: misc_deregister failed");
}
Commit	Line	Data
869d81df	1	/*
2402211a	2	* Copyright (C) 2005-2008 Red Hat, Inc. All rights reserved.
869d81df DT	3	*
	4	* This copyrighted material is made available to anyone wishing to use,
	5	* modify, copy, or redistribute it subject to the terms and conditions
e9fc2aa0	6	* of the GNU General Public License version 2.
869d81df DT	7	*/
869d81df DT	8
2402211a	9	#include <linux/fs.h>
869d81df	10	#include <linux/miscdevice.h>
bd01f843	11	#include <linux/poll.h>
2402211a DT	12	#include <linux/dlm.h>
2402211a DT	13	#include <linux/dlm_plock.h>
5a0e3ad6	14	#include <linux/slab.h>
869d81df	15
2402211a DT	16	#include "dlm_internal.h"
2402211a DT	17	#include "lockspace.h"
869d81df DT	18
	19	static spinlock_t ops_lock;
	20	static struct list_head send_list;
	21	static struct list_head recv_list;
	22	static wait_queue_head_t send_wq;
	23	static wait_queue_head_t recv_wq;
	24
	25	struct plock_op {
	26	struct list_head list;
	27	int done;
2402211a	28	struct dlm_plock_info info;
869d81df DT	29	};
869d81df DT	30
586759f0 ME	31	struct plock_xop {
	32	struct plock_op xop;
	33	void *callback;
	34	void *fl;
	35	void *file;
	36	struct file_lock flc;
	37	};
	38
	39
2402211a	40	static inline void set_version(struct dlm_plock_info *info)
869d81df	41	{
2402211a DT	42	info->version[0] = DLM_PLOCK_VERSION_MAJOR;
	43	info->version[1] = DLM_PLOCK_VERSION_MINOR;
	44	info->version[2] = DLM_PLOCK_VERSION_PATCH;
869d81df DT	45	}
869d81df DT	46
2402211a	47	static int check_version(struct dlm_plock_info *info)
869d81df	48	{
2402211a DT	49	if ((DLM_PLOCK_VERSION_MAJOR != info->version[0]) \|\|
	50	(DLM_PLOCK_VERSION_MINOR < info->version[1])) {
	51	log_print("plock device version mismatch: "
869d81df	52	"kernel (%u.%u.%u), user (%u.%u.%u)",
2402211a DT	53	DLM_PLOCK_VERSION_MAJOR,
	54	DLM_PLOCK_VERSION_MINOR,
	55	DLM_PLOCK_VERSION_PATCH,
869d81df DT	56	info->version[0],
	57	info->version[1],
	58	info->version[2]);
	59	return -EINVAL;
	60	}
	61	return 0;
	62	}
	63
	64	static void send_op(struct plock_op *op)
	65	{
	66	set_version(&op->info);
	67	INIT_LIST_HEAD(&op->list);
	68	spin_lock(&ops_lock);
	69	list_add_tail(&op->list, &send_list);
	70	spin_unlock(&ops_lock);
	71	wake_up(&send_wq);
	72	}
	73
2402211a DT	74	int dlm_posix_lock(dlm_lockspace_t lockspace, u64 number, struct file file,
2402211a DT	75	int cmd, struct file_lock *fl)
869d81df	76	{
2402211a	77	struct dlm_ls *ls;
869d81df	78	struct plock_op *op;
586759f0	79	struct plock_xop *xop;
869d81df DT	80	int rv;
869d81df DT	81
2402211a DT	82	ls = dlm_find_lockspace_local(lockspace);
	83	if (!ls)
	84	return -EINVAL;
	85
573c24c4	86	xop = kzalloc(sizeof(*xop), GFP_NOFS);
2402211a DT	87	if (!xop) {
	88	rv = -ENOMEM;
	89	goto out;
	90	}
869d81df	91
586759f0	92	op = &xop->xop;
2402211a	93	op->info.optype = DLM_PLOCK_OP_LOCK;
3a2a9c96	94	op->info.pid = fl->fl_pid;
869d81df DT	95	op->info.ex = (fl->fl_type == F_WRLCK);
869d81df DT	96	op->info.wait = IS_SETLKW(cmd);
2402211a DT	97	op->info.fsid = ls->ls_global_id;
2402211a DT	98	op->info.number = number;
869d81df DT	99	op->info.start = fl->fl_start;
869d81df DT	100	op->info.end = fl->fl_end;
586759f0	101	if (fl->fl_lmops && fl->fl_lmops->fl_grant) {
2066b58b DT	102	/* fl_owner is lockd which doesn't distinguish
	103	processes on the nfs client */
	104	op->info.owner = (__u64) fl->fl_pid;
586759f0 ME	105	xop->callback = fl->fl_lmops->fl_grant;
	106	locks_init_lock(&xop->flc);
	107	locks_copy_lock(&xop->flc, fl);
	108	xop->fl = fl;
	109	xop->file = file;
2066b58b DT	110	} else {
2066b58b DT	111	op->info.owner = (__u64)(long) fl->fl_owner;
586759f0	112	xop->callback = NULL;
2066b58b	113	}
869d81df DT	114
869d81df DT	115	send_op(op);
586759f0 ME	116
	117	if (xop->callback == NULL)
	118	wait_event(recv_wq, (op->done != 0));
2402211a	119	else {
bde74e4b	120	rv = FILE_LOCK_DEFERRED;
2402211a DT	121	goto out;
2402211a DT	122	}
869d81df DT	123
	124	spin_lock(&ops_lock);
	125	if (!list_empty(&op->list)) {
2402211a DT	126	log_error(ls, "dlm_posix_lock: op on list %llx",
2402211a DT	127	(unsigned long long)number);
869d81df DT	128	list_del(&op->list);
	129	}
	130	spin_unlock(&ops_lock);
	131
	132	rv = op->info.rv;
	133
	134	if (!rv) {
	135	if (posix_lock_file_wait(file, fl) < 0)
2402211a DT	136	log_error(ls, "dlm_posix_lock: vfs lock error %llx",
2402211a DT	137	(unsigned long long)number);
869d81df DT	138	}
869d81df DT	139
586759f0	140	kfree(xop);
2402211a DT	141	out:
2402211a DT	142	dlm_put_lockspace(ls);
586759f0 ME	143	return rv;
586759f0 ME	144	}
2402211a	145	EXPORT_SYMBOL_GPL(dlm_posix_lock);
586759f0	146
af901ca1	147	/* Returns failure iff a successful lock operation should be canceled */
2402211a	148	static int dlm_plock_callback(struct plock_op *op)
586759f0 ME	149	{
	150	struct file *file;
	151	struct file_lock *fl;
	152	struct file_lock *flc;
	153	int (notify)(void , void *, int) = NULL;
	154	struct plock_xop xop = (struct plock_xop )op;
	155	int rv = 0;
	156
	157	spin_lock(&ops_lock);
	158	if (!list_empty(&op->list)) {
2402211a DT	159	log_print("dlm_plock_callback: op on list %llx",
2402211a DT	160	(unsigned long long)op->info.number);
586759f0 ME	161	list_del(&op->list);
	162	}
	163	spin_unlock(&ops_lock);
	164
	165	/* check if the following 2 are still valid or make a copy */
	166	file = xop->file;
	167	flc = &xop->flc;
	168	fl = xop->fl;
	169	notify = xop->callback;
	170
	171	if (op->info.rv) {
24179f48	172	notify(fl, NULL, op->info.rv);
586759f0 ME	173	goto out;
	174	}
	175
	176	/* got fs lock; bookkeep locally as well: */
	177	flc->fl_flags &= ~FL_SLEEP;
	178	if (posix_lock_file(file, flc, NULL)) {
	179	/*
	180	* This can only happen in the case of kmalloc() failure.
	181	* The filesystem's own lock is the authoritative lock,
	182	* so a failure to get the lock locally is not a disaster.
2402211a	183	* As long as the fs cannot reliably cancel locks (especially
586759f0 ME	184	* in a low-memory situation), we're better off ignoring
	185	* this failure than trying to recover.
	186	*/
2402211a DT	187	log_print("dlm_plock_callback: vfs lock error %llx file %p fl %p",
2402211a DT	188	(unsigned long long)op->info.number, file, fl);
586759f0 ME	189	}
586759f0 ME	190
24179f48	191	rv = notify(fl, NULL, 0);
586759f0 ME	192	if (rv) {
586759f0 ME	193	/* XXX: We need to cancel the fs lock here: */
2402211a DT	194	log_print("dlm_plock_callback: lock granted after lock request "
2402211a DT	195	"failed; dangling lock!\n");
586759f0 ME	196	goto out;
	197	}
	198
	199	out:
	200	kfree(xop);
869d81df DT	201	return rv;
	202	}
	203
2402211a DT	204	int dlm_posix_unlock(dlm_lockspace_t lockspace, u64 number, struct file file,
2402211a DT	205	struct file_lock *fl)
869d81df	206	{
2402211a	207	struct dlm_ls *ls;
869d81df DT	208	struct plock_op *op;
	209	int rv;
	210
2402211a DT	211	ls = dlm_find_lockspace_local(lockspace);
	212	if (!ls)
	213	return -EINVAL;
	214
573c24c4	215	op = kzalloc(sizeof(*op), GFP_NOFS);
2402211a DT	216	if (!op) {
	217	rv = -ENOMEM;
	218	goto out;
	219	}
869d81df DT	220
869d81df DT	221	if (posix_lock_file_wait(file, fl) < 0)
2402211a DT	222	log_error(ls, "dlm_posix_unlock: vfs unlock error %llx",
2402211a DT	223	(unsigned long long)number);
869d81df	224
2402211a	225	op->info.optype = DLM_PLOCK_OP_UNLOCK;
3a2a9c96	226	op->info.pid = fl->fl_pid;
2402211a DT	227	op->info.fsid = ls->ls_global_id;
2402211a DT	228	op->info.number = number;
869d81df DT	229	op->info.start = fl->fl_start;
869d81df DT	230	op->info.end = fl->fl_end;
2066b58b DT	231	if (fl->fl_lmops && fl->fl_lmops->fl_grant)
	232	op->info.owner = (__u64) fl->fl_pid;
	233	else
	234	op->info.owner = (__u64)(long) fl->fl_owner;
869d81df DT	235
	236	send_op(op);
	237	wait_event(recv_wq, (op->done != 0));
	238
	239	spin_lock(&ops_lock);
	240	if (!list_empty(&op->list)) {
2402211a DT	241	log_error(ls, "dlm_posix_unlock: op on list %llx",
2402211a DT	242	(unsigned long long)number);
869d81df DT	243	list_del(&op->list);
	244	}
	245	spin_unlock(&ops_lock);
	246
	247	rv = op->info.rv;
	248
586759f0 ME	249	if (rv == -ENOENT)
	250	rv = 0;
	251
869d81df	252	kfree(op);
2402211a DT	253	out:
2402211a DT	254	dlm_put_lockspace(ls);
869d81df DT	255	return rv;
869d81df DT	256	}
2402211a	257	EXPORT_SYMBOL_GPL(dlm_posix_unlock);
869d81df	258
2402211a DT	259	int dlm_posix_get(dlm_lockspace_t lockspace, u64 number, struct file file,
2402211a DT	260	struct file_lock *fl)
869d81df	261	{
2402211a	262	struct dlm_ls *ls;
869d81df DT	263	struct plock_op *op;
	264	int rv;
	265
2402211a DT	266	ls = dlm_find_lockspace_local(lockspace);
	267	if (!ls)
	268	return -EINVAL;
	269
573c24c4	270	op = kzalloc(sizeof(*op), GFP_NOFS);
2402211a DT	271	if (!op) {
	272	rv = -ENOMEM;
	273	goto out;
	274	}
869d81df	275
2402211a	276	op->info.optype = DLM_PLOCK_OP_GET;
3a2a9c96	277	op->info.pid = fl->fl_pid;
869d81df	278	op->info.ex = (fl->fl_type == F_WRLCK);
2402211a DT	279	op->info.fsid = ls->ls_global_id;
2402211a DT	280	op->info.number = number;
869d81df DT	281	op->info.start = fl->fl_start;
869d81df DT	282	op->info.end = fl->fl_end;
2066b58b DT	283	if (fl->fl_lmops && fl->fl_lmops->fl_grant)
	284	op->info.owner = (__u64) fl->fl_pid;
	285	else
	286	op->info.owner = (__u64)(long) fl->fl_owner;
586759f0	287
869d81df DT	288	send_op(op);
	289	wait_event(recv_wq, (op->done != 0));
	290
	291	spin_lock(&ops_lock);
	292	if (!list_empty(&op->list)) {
2402211a DT	293	log_error(ls, "dlm_posix_get: op on list %llx",
2402211a DT	294	(unsigned long long)number);
869d81df DT	295	list_del(&op->list);
	296	}
	297	spin_unlock(&ops_lock);
	298
a7a2ff8a DT	299	/* info.rv from userspace is 1 for conflict, 0 for no-conflict,
	300	-ENOENT if there are no locks on the file */
	301
869d81df DT	302	rv = op->info.rv;
869d81df DT	303
586759f0 ME	304	fl->fl_type = F_UNLCK;
	305	if (rv == -ENOENT)
	306	rv = 0;
a7a2ff8a	307	else if (rv > 0) {
20d5a399	308	locks_init_lock(fl);
869d81df	309	fl->fl_type = (op->info.ex) ? F_WRLCK : F_RDLCK;
20d5a399	310	fl->fl_flags = FL_POSIX;
869d81df DT	311	fl->fl_pid = op->info.pid;
	312	fl->fl_start = op->info.start;
	313	fl->fl_end = op->info.end;
a7a2ff8a	314	rv = 0;
869d81df DT	315	}
	316
	317	kfree(op);
2402211a DT	318	out:
2402211a DT	319	dlm_put_lockspace(ls);
869d81df DT	320	return rv;
869d81df DT	321	}
2402211a	322	EXPORT_SYMBOL_GPL(dlm_posix_get);
869d81df DT	323
	324	/* a read copies out one plock request from the send list */
	325	static ssize_t dev_read(struct file file, char __user u, size_t count,
	326	loff_t *ppos)
	327	{
2402211a	328	struct dlm_plock_info info;
869d81df DT	329	struct plock_op *op = NULL;
	330
	331	if (count < sizeof(info))
	332	return -EINVAL;
	333
	334	spin_lock(&ops_lock);
	335	if (!list_empty(&send_list)) {
	336	op = list_entry(send_list.next, struct plock_op, list);
	337	list_move(&op->list, &recv_list);
	338	memcpy(&info, &op->info, sizeof(info));
	339	}
	340	spin_unlock(&ops_lock);
	341
	342	if (!op)
	343	return -EAGAIN;
	344
	345	if (copy_to_user(u, &info, sizeof(info)))
	346	return -EFAULT;
	347	return sizeof(info);
	348	}
	349
	350	/* a write copies in one plock result that should match a plock_op
	351	on the recv list */
	352	static ssize_t dev_write(struct file file, const char __user u, size_t count,
	353	loff_t *ppos)
	354	{
2402211a	355	struct dlm_plock_info info;
869d81df	356	struct plock_op *op;
c78a87d0	357	int found = 0, do_callback = 0;
869d81df DT	358
	359	if (count != sizeof(info))
	360	return -EINVAL;
	361
	362	if (copy_from_user(&info, u, sizeof(info)))
	363	return -EFAULT;
	364
	365	if (check_version(&info))
	366	return -EINVAL;
	367
	368	spin_lock(&ops_lock);
	369	list_for_each_entry(op, &recv_list, list) {
c78a87d0 DT	370	if (op->info.fsid == info.fsid &&
c78a87d0 DT	371	op->info.number == info.number &&
08eac93a	372	op->info.owner == info.owner) {
c78a87d0	373	struct plock_xop xop = (struct plock_xop )op;
869d81df	374	list_del_init(&op->list);
869d81df	375	memcpy(&op->info, &info, sizeof(info));
c78a87d0 DT	376	if (xop->callback)
	377	do_callback = 1;
	378	else
	379	op->done = 1;
	380	found = 1;
869d81df DT	381	break;
	382	}
	383	}
	384	spin_unlock(&ops_lock);
	385
586759f0	386	if (found) {
c78a87d0	387	if (do_callback)
817d10ba	388	dlm_plock_callback(op);
586759f0 ME	389	else
	390	wake_up(&recv_wq);
	391	} else
2402211a DT	392	log_print("dev_write no op %x %llx", info.fsid,
2402211a DT	393	(unsigned long long)info.number);
869d81df DT	394	return count;
	395	}
	396
	397	static unsigned int dev_poll(struct file file, poll_table wait)
	398	{
cee23c79 DC	399	unsigned int mask = 0;
cee23c79 DC	400
869d81df DT	401	poll_wait(file, &send_wq, wait);
	402
	403	spin_lock(&ops_lock);
cee23c79 DC	404	if (!list_empty(&send_list))
cee23c79 DC	405	mask = POLLIN \| POLLRDNORM;
869d81df	406	spin_unlock(&ops_lock);
cee23c79 DC	407
cee23c79 DC	408	return mask;
869d81df DT	409	}
869d81df DT	410
00977a59	411	static const struct file_operations dev_fops = {
869d81df DT	412	.read = dev_read,
	413	.write = dev_write,
	414	.poll = dev_poll,
6038f373 AB	415	.owner = THIS_MODULE,
6038f373 AB	416	.llseek = noop_llseek,
869d81df DT	417	};
	418
	419	static struct miscdevice plock_dev_misc = {
	420	.minor = MISC_DYNAMIC_MINOR,
2402211a	421	.name = DLM_PLOCK_MISC_NAME,
869d81df DT	422	.fops = &dev_fops
	423	};
	424
2402211a	425	int dlm_plock_init(void)
869d81df DT	426	{
	427	int rv;
	428
	429	spin_lock_init(&ops_lock);
	430	INIT_LIST_HEAD(&send_list);
	431	INIT_LIST_HEAD(&recv_list);
	432	init_waitqueue_head(&send_wq);
	433	init_waitqueue_head(&recv_wq);
	434
	435	rv = misc_register(&plock_dev_misc);
	436	if (rv)
2402211a	437	log_print("dlm_plock_init: misc_register failed %d", rv);
869d81df DT	438	return rv;
	439	}
	440
2402211a	441	void dlm_plock_exit(void)
869d81df DT	442	{
869d81df DT	443	if (misc_deregister(&plock_dev_misc) < 0)
2402211a	444	log_print("dlm_plock_exit: misc_deregister failed");
869d81df DT	445	}
869d81df DT	446