[net-next-2.6.git] / drivers / block / aoe / aoedev.c

/* Copyright (c) 2007 Coraid, Inc.  See COPYING for GPL terms. */
/*
 * aoedev.c
 * AoE device utility functions; maintains device list.
 */

#include <linux/hdreg.h>
#include <linux/blkdev.h>
#include <linux/netdevice.h>
#include <linux/delay.h>
#include "aoe.h"

static void dummy_timer(ulong);
static void aoedev_freedev(struct aoedev *);
static void freetgt(struct aoedev *d, struct aoetgt *t);
static void skbpoolfree(struct aoedev *d);

static struct aoedev *devlist;
static DEFINE_SPINLOCK(devlist_lock);

struct aoedev *
aoedev_by_aoeaddr(int maj, int min)
{
	struct aoedev *d;
	ulong flags;

	spin_lock_irqsave(&devlist_lock, flags);

	for (d=devlist; d; d=d->next)
		if (d->aoemajor == maj && d->aoeminor == min)
			break;

	spin_unlock_irqrestore(&devlist_lock, flags);
	return d;
}

static void
dummy_timer(ulong vp)
{
	struct aoedev *d;

	d = (struct aoedev *)vp;
	if (d->flags & DEVFL_TKILL)
		return;
	d->timer.expires = jiffies + HZ;
	add_timer(&d->timer);
}

void
aoedev_downdev(struct aoedev *d)
{
	struct aoetgt **t, **te;
	struct frame *f, *e;
	struct buf *buf;
	struct bio *bio;

	t = d->targets;
	te = t + NTARGETS;
	for (; t < te && *t; t++) {
		f = (*t)->frames;
		e = f + (*t)->nframes;
		for (; f < e; f->tag = FREETAG, f->buf = NULL, f++) {
			if (f->tag == FREETAG || f->buf == NULL)
				continue;
			buf = f->buf;
			bio = buf->bio;
			if (--buf->nframesout == 0
			&& buf != d->inprocess) {
				mempool_free(buf, d->bufpool);
				bio_endio(bio, -EIO);
			}
		}
		(*t)->maxout = (*t)->nframes;
		(*t)->nout = 0;
	}
	buf = d->inprocess;
	if (buf) {
		bio = buf->bio;
		mempool_free(buf, d->bufpool);
		bio_endio(bio, -EIO);
	}
	d->inprocess = NULL;
	d->htgt = NULL;

	while (!list_empty(&d->bufq)) {
		buf = container_of(d->bufq.next, struct buf, bufs);
		list_del(d->bufq.next);
		bio = buf->bio;
		mempool_free(buf, d->bufpool);
		bio_endio(bio, -EIO);
	}

	if (d->gd)
		set_capacity(d->gd, 0);

	d->flags &= ~DEVFL_UP;
}

static void
aoedev_freedev(struct aoedev *d)
{
	struct aoetgt **t, **e;

	if (d->gd) {
		aoedisk_rm_sysfs(d);
		del_gendisk(d->gd);
		put_disk(d->gd);
	}
	t = d->targets;
	e = t + NTARGETS;
	for (; t < e && *t; t++)
		freetgt(d, *t);
	if (d->bufpool)
		mempool_destroy(d->bufpool);
	skbpoolfree(d);
	kfree(d);
}

int
aoedev_flush(const char __user *str, size_t cnt)
{
	ulong flags;
	struct aoedev *d, **dd;
	struct aoedev *rmd = NULL;
	char buf[16];
	int all = 0;

	if (cnt >= 3) {
		if (cnt > sizeof buf)
			cnt = sizeof buf;
		if (copy_from_user(buf, str, cnt))
			return -EFAULT;
		all = !strncmp(buf, "all", 3);
	}

	flush_scheduled_work();
	spin_lock_irqsave(&devlist_lock, flags);
	dd = &devlist;
	while ((d = *dd)) {
		spin_lock(&d->lock);
		if ((!all && (d->flags & DEVFL_UP))
		|| (d->flags & (DEVFL_GDALLOC|DEVFL_NEWSIZE))
		|| d->nopen) {
			spin_unlock(&d->lock);
			dd = &d->next;
			continue;
		}
		*dd = d->next;
		aoedev_downdev(d);
		d->flags |= DEVFL_TKILL;
		spin_unlock(&d->lock);
		d->next = rmd;
		rmd = d;
	}
	spin_unlock_irqrestore(&devlist_lock, flags);
	while ((d = rmd)) {
		rmd = d->next;
		del_timer_sync(&d->timer);
		aoedev_freedev(d);	/* must be able to sleep */
	}
	return 0;
}

/* I'm not really sure that this is a realistic problem, but if the
network driver goes gonzo let's just leak memory after complaining. */
static void
skbfree(struct sk_buff *skb)
{
	enum { Sms = 100, Tms = 3*1000};
	int i = Tms / Sms;

	if (skb == NULL)
		return;
	while (atomic_read(&skb_shinfo(skb)->dataref) != 1 && i-- > 0)
		msleep(Sms);
	if (i <= 0) {
		printk(KERN_ERR
			"aoe: %s holds ref: %s\n",
			skb->dev ? skb->dev->name : "netif",
			"cannot free skb -- memory leaked.");
		return;
	}
	skb_shinfo(skb)->nr_frags = skb->data_len = 0;
	skb_trim(skb, 0);
	dev_kfree_skb(skb);
}

static void
skbpoolfree(struct aoedev *d)
{
	struct sk_buff *skb, *tmp;

	skb_queue_walk_safe(&d->skbpool, skb, tmp)
		skbfree(skb);

	__skb_queue_head_init(&d->skbpool);
}

/* find it or malloc it */
struct aoedev *
aoedev_by_sysminor_m(ulong sysminor)
{
	struct aoedev *d;
	ulong flags;

	spin_lock_irqsave(&devlist_lock, flags);

	for (d=devlist; d; d=d->next)
		if (d->sysminor == sysminor)
			break;
	if (d)
		goto out;
	d = kcalloc(1, sizeof *d, GFP_ATOMIC);
	if (!d)
		goto out;
	INIT_WORK(&d->work, aoecmd_sleepwork);
	spin_lock_init(&d->lock);
	skb_queue_head_init(&d->sendq);
	skb_queue_head_init(&d->skbpool);
	init_timer(&d->timer);
	d->timer.data = (ulong) d;
	d->timer.function = dummy_timer;
	d->timer.expires = jiffies + HZ;
	add_timer(&d->timer);
	d->bufpool = NULL;	/* defer to aoeblk_gdalloc */
	d->tgt = d->targets;
	INIT_LIST_HEAD(&d->bufq);
	d->sysminor = sysminor;
	d->aoemajor = AOEMAJOR(sysminor);
	d->aoeminor = AOEMINOR(sysminor);
	d->mintimer = MINTIMER;
	d->next = devlist;
	devlist = d;
 out:
	spin_unlock_irqrestore(&devlist_lock, flags);
	return d;
}

static void
freetgt(struct aoedev *d, struct aoetgt *t)
{
	struct frame *f, *e;

	f = t->frames;
	e = f + t->nframes;
	for (; f < e; f++)
		skbfree(f->skb);
	kfree(t->frames);
	kfree(t);
}

void
aoedev_exit(void)
{
	struct aoedev *d;
	ulong flags;

	flush_scheduled_work();

	while ((d = devlist)) {
		devlist = d->next;

		spin_lock_irqsave(&d->lock, flags);
		aoedev_downdev(d);
		d->flags |= DEVFL_TKILL;
		spin_unlock_irqrestore(&d->lock, flags);

		del_timer_sync(&d->timer);
		aoedev_freedev(d);
	}
}

int __init
aoedev_init(void)
{
	return 0;
}
Commit	Line	Data
52e112b3	1	/* Copyright (c) 2007 Coraid, Inc. See COPYING for GPL terms. */
1da177e4 LT	2	/*
	3	* aoedev.c
	4	* AoE device utility functions; maintains device list.
	5	*/
	6
	7	#include <linux/hdreg.h>
	8	#include <linux/blkdev.h>
	9	#include <linux/netdevice.h>
9bb237b6	10	#include <linux/delay.h>
1da177e4 LT	11	#include "aoe.h"
1da177e4 LT	12
262bf541 EC	13	static void dummy_timer(ulong);
262bf541 EC	14	static void aoedev_freedev(struct aoedev *);
9bb237b6 EC	15	static void freetgt(struct aoedev d, struct aoetgt t);
9bb237b6 EC	16	static void skbpoolfree(struct aoedev *d);
262bf541	17
1da177e4	18	static struct aoedev *devlist;
476aed38	19	static DEFINE_SPINLOCK(devlist_lock);
1da177e4 LT	20
1da177e4 LT	21	struct aoedev *
32465c65	22	aoedev_by_aoeaddr(int maj, int min)
1da177e4 LT	23	{
	24	struct aoedev *d;
	25	ulong flags;
	26
	27	spin_lock_irqsave(&devlist_lock, flags);
	28
	29	for (d=devlist; d; d=d->next)
32465c65	30	if (d->aoemajor == maj && d->aoeminor == min)
1da177e4 LT	31	break;
	32
	33	spin_unlock_irqrestore(&devlist_lock, flags);
	34	return d;
	35	}
	36
3ae1c24e EC	37	static void
	38	dummy_timer(ulong vp)
	39	{
	40	struct aoedev *d;
	41
	42	d = (struct aoedev *)vp;
	43	if (d->flags & DEVFL_TKILL)
	44	return;
	45	d->timer.expires = jiffies + HZ;
	46	add_timer(&d->timer);
	47	}
	48
1da177e4 LT	49	void
	50	aoedev_downdev(struct aoedev *d)
	51	{
68e0d42f	52	struct aoetgt t, te;
1da177e4 LT	53	struct frame f, e;
	54	struct buf *buf;
	55	struct bio *bio;
	56
68e0d42f EC	57	t = d->targets;
	58	te = t + NTARGETS;
	59	for (; t < te && *t; t++) {
	60	f = (*t)->frames;
	61	e = f + (*t)->nframes;
	62	for (; f < e; f->tag = FREETAG, f->buf = NULL, f++) {
	63	if (f->tag == FREETAG \|\| f->buf == NULL)
	64	continue;
	65	buf = f->buf;
	66	bio = buf->bio;
	67	if (--buf->nframesout == 0
	68	&& buf != d->inprocess) {
	69	mempool_free(buf, d->bufpool);
	70	bio_endio(bio, -EIO);
	71	}
1da177e4	72	}
68e0d42f EC	73	(t)->maxout = (t)->nframes;
	74	(*t)->nout = 0;
	75	}
	76	buf = d->inprocess;
	77	if (buf) {
	78	bio = buf->bio;
	79	mempool_free(buf, d->bufpool);
	80	bio_endio(bio, -EIO);
1da177e4 LT	81	}
1da177e4 LT	82	d->inprocess = NULL;
68e0d42f	83	d->htgt = NULL;
1da177e4 LT	84
	85	while (!list_empty(&d->bufq)) {
	86	buf = container_of(d->bufq.next, struct buf, bufs);
	87	list_del(d->bufq.next);
	88	bio = buf->bio;
	89	mempool_free(buf, d->bufpool);
6712ecf8	90	bio_endio(bio, -EIO);
1da177e4 LT	91	}
1da177e4 LT	92
1da177e4	93	if (d->gd)
80795aef	94	set_capacity(d->gd, 0);
1da177e4	95
68e0d42f	96	d->flags &= ~DEVFL_UP;
1da177e4 LT	97	}
1da177e4 LT	98
262bf541 EC	99	static void
	100	aoedev_freedev(struct aoedev *d)
	101	{
	102	struct aoetgt t, e;
	103
	104	if (d->gd) {
	105	aoedisk_rm_sysfs(d);
	106	del_gendisk(d->gd);
	107	put_disk(d->gd);
	108	}
	109	t = d->targets;
	110	e = t + NTARGETS;
	111	for (; t < e && *t; t++)
9bb237b6	112	freetgt(d, *t);
262bf541 EC	113	if (d->bufpool)
262bf541 EC	114	mempool_destroy(d->bufpool);
9bb237b6	115	skbpoolfree(d);
262bf541 EC	116	kfree(d);
	117	}
	118
	119	int
	120	aoedev_flush(const char __user *str, size_t cnt)
	121	{
	122	ulong flags;
	123	struct aoedev d, *dd;
	124	struct aoedev *rmd = NULL;
	125	char buf[16];
	126	int all = 0;
	127
	128	if (cnt >= 3) {
	129	if (cnt > sizeof buf)
	130	cnt = sizeof buf;
	131	if (copy_from_user(buf, str, cnt))
	132	return -EFAULT;
	133	all = !strncmp(buf, "all", 3);
	134	}
	135
	136	flush_scheduled_work();
	137	spin_lock_irqsave(&devlist_lock, flags);
	138	dd = &devlist;
	139	while ((d = *dd)) {
	140	spin_lock(&d->lock);
	141	if ((!all && (d->flags & DEVFL_UP))
	142	\|\| (d->flags & (DEVFL_GDALLOC\|DEVFL_NEWSIZE))
	143	\|\| d->nopen) {
	144	spin_unlock(&d->lock);
	145	dd = &d->next;
	146	continue;
	147	}
	148	*dd = d->next;
	149	aoedev_downdev(d);
	150	d->flags \|= DEVFL_TKILL;
	151	spin_unlock(&d->lock);
	152	d->next = rmd;
	153	rmd = d;
	154	}
	155	spin_unlock_irqrestore(&devlist_lock, flags);
	156	while ((d = rmd)) {
	157	rmd = d->next;
	158	del_timer_sync(&d->timer);
	159	aoedev_freedev(d); /* must be able to sleep */
	160	}
	161	return 0;
	162	}
	163
9bb237b6 EC	164	/* I'm not really sure that this is a realistic problem, but if the
	165	network driver goes gonzo let's just leak memory after complaining. */
	166	static void
	167	skbfree(struct sk_buff *skb)
	168	{
	169	enum { Sms = 100, Tms = 3*1000};
	170	int i = Tms / Sms;
	171
	172	if (skb == NULL)
	173	return;
	174	while (atomic_read(&skb_shinfo(skb)->dataref) != 1 && i-- > 0)
	175	msleep(Sms);
	176	if (i <= 0) {
	177	printk(KERN_ERR
	178	"aoe: %s holds ref: %s\n",
	179	skb->dev ? skb->dev->name : "netif",
	180	"cannot free skb -- memory leaked.");
	181	return;
	182	}
	183	skb_shinfo(skb)->nr_frags = skb->data_len = 0;
	184	skb_trim(skb, 0);
	185	dev_kfree_skb(skb);
	186	}
	187
	188	static void
	189	skbpoolfree(struct aoedev *d)
	190	{
e9bb8fb0	191	struct sk_buff skb, tmp;
9bb237b6	192
e9bb8fb0	193	skb_queue_walk_safe(&d->skbpool, skb, tmp)
9bb237b6	194	skbfree(skb);
e9bb8fb0 DM	195
e9bb8fb0 DM	196	__skb_queue_head_init(&d->skbpool);
9bb237b6 EC	197	}
9bb237b6 EC	198
3ae1c24e	199	/* find it or malloc it */
1da177e4	200	struct aoedev *
68e0d42f	201	aoedev_by_sysminor_m(ulong sysminor)
1da177e4 LT	202	{
	203	struct aoedev *d;
	204	ulong flags;
	205
	206	spin_lock_irqsave(&devlist_lock, flags);
	207
	208	for (d=devlist; d; d=d->next)
93d489fc	209	if (d->sysminor == sysminor)
1da177e4	210	break;
68e0d42f EC	211	if (d)
	212	goto out;
	213	d = kcalloc(1, sizeof *d, GFP_ATOMIC);
	214	if (!d)
	215	goto out;
	216	INIT_WORK(&d->work, aoecmd_sleepwork);
	217	spin_lock_init(&d->lock);
e9bb8fb0 DM	218	skb_queue_head_init(&d->sendq);
e9bb8fb0 DM	219	skb_queue_head_init(&d->skbpool);
68e0d42f EC	220	init_timer(&d->timer);
	221	d->timer.data = (ulong) d;
	222	d->timer.function = dummy_timer;
	223	d->timer.expires = jiffies + HZ;
	224	add_timer(&d->timer);
	225	d->bufpool = NULL; /* defer to aoeblk_gdalloc */
	226	d->tgt = d->targets;
	227	INIT_LIST_HEAD(&d->bufq);
	228	d->sysminor = sysminor;
	229	d->aoemajor = AOEMAJOR(sysminor);
	230	d->aoeminor = AOEMINOR(sysminor);
	231	d->mintimer = MINTIMER;
	232	d->next = devlist;
	233	devlist = d;
	234	out:
3ae1c24e	235	spin_unlock_irqrestore(&devlist_lock, flags);
1da177e4 LT	236	return d;
	237	}
	238
	239	static void
9bb237b6	240	freetgt(struct aoedev d, struct aoetgt t)
1da177e4	241	{
e407a7f6 EC	242	struct frame f, e;
e407a7f6 EC	243
68e0d42f EC	244	f = t->frames;
68e0d42f EC	245	e = f + t->nframes;
9bb237b6 EC	246	for (; f < e; f++)
9bb237b6 EC	247	skbfree(f->skb);
68e0d42f EC	248	kfree(t->frames);
	249	kfree(t);
	250	}
	251
1da177e4 LT	252	void
	253	aoedev_exit(void)
	254	{
	255	struct aoedev *d;
	256	ulong flags;
	257
	258	flush_scheduled_work();
	259
	260	while ((d = devlist)) {
	261	devlist = d->next;
	262
	263	spin_lock_irqsave(&d->lock, flags);
	264	aoedev_downdev(d);
3ae1c24e	265	d->flags \|= DEVFL_TKILL;
1da177e4 LT	266	spin_unlock_irqrestore(&d->lock, flags);
	267
	268	del_timer_sync(&d->timer);
	269	aoedev_freedev(d);
	270	}
	271	}
	272
	273	int __init
	274	aoedev_init(void)
	275	{
1da177e4 LT	276	return 0;
1da177e4 LT	277	}