[linux-2.6-block.git] / fs / orangefs / orangefs-bufmap.c

// SPDX-License-Identifier: GPL-2.0
/*
 * (C) 2001 Clemson University and The University of Chicago
 *
 * See COPYING in top-level directory.
 */
#include "protocol.h"
#include "orangefs-kernel.h"
#include "orangefs-bufmap.h"

struct slot_map {
	int c;
	wait_queue_head_t q;
	int count;
	unsigned long *map;
};

static struct slot_map rw_map = {
	.c = -1,
	.q = __WAIT_QUEUE_HEAD_INITIALIZER(rw_map.q)
};
static struct slot_map readdir_map = {
	.c = -1,
	.q = __WAIT_QUEUE_HEAD_INITIALIZER(readdir_map.q)
};


static void install(struct slot_map *m, int count, unsigned long *map)
{
	spin_lock(&m->q.lock);
	m->c = m->count = count;
	m->map = map;
	wake_up_all_locked(&m->q);
	spin_unlock(&m->q.lock);
}

static void mark_killed(struct slot_map *m)
{
	spin_lock(&m->q.lock);
	m->c -= m->count + 1;
	spin_unlock(&m->q.lock);
}

static void run_down(struct slot_map *m)
{
	DEFINE_WAIT(wait);
	spin_lock(&m->q.lock);
	if (m->c != -1) {
		for (;;) {
			if (likely(list_empty(&wait.entry)))
				__add_wait_queue_entry_tail(&m->q, &wait);
			set_current_state(TASK_UNINTERRUPTIBLE);

			if (m->c == -1)
				break;

			spin_unlock(&m->q.lock);
			schedule();
			spin_lock(&m->q.lock);
		}
		__remove_wait_queue(&m->q, &wait);
		__set_current_state(TASK_RUNNING);
	}
	m->map = NULL;
	spin_unlock(&m->q.lock);
}

static void put(struct slot_map *m, int slot)
{
	int v;
	spin_lock(&m->q.lock);
	__clear_bit(slot, m->map);
	v = ++m->c;
	if (unlikely(v == 1))	/* no free slots -> one free slot */
		wake_up_locked(&m->q);
	else if (unlikely(v == -1))	/* finished dying */
		wake_up_all_locked(&m->q);
	spin_unlock(&m->q.lock);
}

static int wait_for_free(struct slot_map *m)
{
	long left = slot_timeout_secs * HZ;
	DEFINE_WAIT(wait);

	do {
		long n = left, t;
		if (likely(list_empty(&wait.entry)))
			__add_wait_queue_entry_tail_exclusive(&m->q, &wait);
		set_current_state(TASK_INTERRUPTIBLE);

		if (m->c > 0)
			break;

		if (m->c < 0) {
			/* we are waiting for map to be installed */
			/* it would better be there soon, or we go away */
			if (n > ORANGEFS_BUFMAP_WAIT_TIMEOUT_SECS * HZ)
				n = ORANGEFS_BUFMAP_WAIT_TIMEOUT_SECS * HZ;
		}
		spin_unlock(&m->q.lock);
		t = schedule_timeout(n);
		spin_lock(&m->q.lock);
		if (unlikely(!t) && n != left && m->c < 0)
			left = t;
		else
			left = t + (left - n);
		if (unlikely(signal_pending(current)))
			left = -EINTR;
	} while (left > 0);

	if (!list_empty(&wait.entry))
		list_del(&wait.entry);
	else if (left <= 0 && waitqueue_active(&m->q))
		__wake_up_locked_key(&m->q, TASK_INTERRUPTIBLE, NULL);
	__set_current_state(TASK_RUNNING);

	if (likely(left > 0))
		return 0;

	return left < 0 ? -EINTR : -ETIMEDOUT;
}

static int get(struct slot_map *m)
{
	int res = 0;
	spin_lock(&m->q.lock);
	if (unlikely(m->c <= 0))
		res = wait_for_free(m);
	if (likely(!res)) {
		m->c--;
		res = find_first_zero_bit(m->map, m->count);
		__set_bit(res, m->map);
	}
	spin_unlock(&m->q.lock);
	return res;
}

/* used to describe mapped buffers */
struct orangefs_bufmap_desc {
	void *uaddr;			/* user space address pointer */
	struct page **page_array;	/* array of mapped pages */
	int array_count;		/* size of above arrays */
	struct list_head list_link;
};

static struct orangefs_bufmap {
	int desc_size;
	int desc_shift;
	int desc_count;
	int total_size;
	int page_count;

	struct page **page_array;
	struct orangefs_bufmap_desc *desc_array;

	/* array to track usage of buffer descriptors */
	unsigned long *buffer_index_array;

	/* array to track usage of buffer descriptors for readdir */
#define N DIV_ROUND_UP(ORANGEFS_READDIR_DEFAULT_DESC_COUNT, BITS_PER_LONG)
	unsigned long readdir_index_array[N];
#undef N
} *__orangefs_bufmap;

static DEFINE_SPINLOCK(orangefs_bufmap_lock);

static void
orangefs_bufmap_unmap(struct orangefs_bufmap *bufmap)
{
	int i;

	for (i = 0; i < bufmap->page_count; i++)
		put_page(bufmap->page_array[i]);
}

static void
orangefs_bufmap_free(struct orangefs_bufmap *bufmap)
{
	kfree(bufmap->page_array);
	kfree(bufmap->desc_array);
	kfree(bufmap->buffer_index_array);
	kfree(bufmap);
}

/*
 * XXX: Can the size and shift change while the caller gives up the 
 * XXX: lock between calling this and doing something useful?
 */

int orangefs_bufmap_size_query(void)
{
	struct orangefs_bufmap *bufmap;
	int size = 0;
	spin_lock(&orangefs_bufmap_lock);
	bufmap = __orangefs_bufmap;
	if (bufmap)
		size = bufmap->desc_size;
	spin_unlock(&orangefs_bufmap_lock);
	return size;
}

int orangefs_bufmap_shift_query(void)
{
	struct orangefs_bufmap *bufmap;
	int shift = 0;
	spin_lock(&orangefs_bufmap_lock);
	bufmap = __orangefs_bufmap;
	if (bufmap)
		shift = bufmap->desc_shift;
	spin_unlock(&orangefs_bufmap_lock);
	return shift;
}

static DECLARE_WAIT_QUEUE_HEAD(bufmap_waitq);
static DECLARE_WAIT_QUEUE_HEAD(readdir_waitq);

/*
 * orangefs_get_bufmap_init
 *
 * If bufmap_init is 1, then the shared memory system, including the
 * buffer_index_array, is available.  Otherwise, it is not.
 *
 * returns the value of bufmap_init
 */
int orangefs_get_bufmap_init(void)
{
	return __orangefs_bufmap ? 1 : 0;
}


static struct orangefs_bufmap *
orangefs_bufmap_alloc(struct ORANGEFS_dev_map_desc *user_desc)
{
	struct orangefs_bufmap *bufmap;

	bufmap = kzalloc(sizeof(*bufmap), GFP_KERNEL);
	if (!bufmap)
		goto out;

	bufmap->total_size = user_desc->total_size;
	bufmap->desc_count = user_desc->count;
	bufmap->desc_size = user_desc->size;
	bufmap->desc_shift = ilog2(bufmap->desc_size);

	bufmap->buffer_index_array =
		kzalloc(DIV_ROUND_UP(bufmap->desc_count, BITS_PER_LONG), GFP_KERNEL);
	if (!bufmap->buffer_index_array)
		goto out_free_bufmap;

	bufmap->desc_array =
		kcalloc(bufmap->desc_count, sizeof(struct orangefs_bufmap_desc),
			GFP_KERNEL);
	if (!bufmap->desc_array)
		goto out_free_index_array;

	bufmap->page_count = bufmap->total_size / PAGE_SIZE;

	/* allocate storage to track our page mappings */
	bufmap->page_array =
		kcalloc(bufmap->page_count, sizeof(struct page *), GFP_KERNEL);
	if (!bufmap->page_array)
		goto out_free_desc_array;

	return bufmap;

out_free_desc_array:
	kfree(bufmap->desc_array);
out_free_index_array:
	kfree(bufmap->buffer_index_array);
out_free_bufmap:
	kfree(bufmap);
out:
	return NULL;
}

static int
orangefs_bufmap_map(struct orangefs_bufmap *bufmap,
		struct ORANGEFS_dev_map_desc *user_desc)
{
	int pages_per_desc = bufmap->desc_size / PAGE_SIZE;
	int offset = 0, ret, i;

	/* map the pages */
	ret = get_user_pages_fast((unsigned long)user_desc->ptr,
			     bufmap->page_count, 1, bufmap->page_array);

	if (ret < 0)
		return ret;

	if (ret != bufmap->page_count) {
		gossip_err("orangefs error: asked for %d pages, only got %d.\n",
				bufmap->page_count, ret);

		for (i = 0; i < ret; i++) {
			SetPageError(bufmap->page_array[i]);
			put_page(bufmap->page_array[i]);
		}
		return -ENOMEM;
	}

	/*
	 * ideally we want to get kernel space pointers for each page, but
	 * we can't kmap that many pages at once if highmem is being used.
	 * so instead, we just kmap/kunmap the page address each time the
	 * kaddr is needed.
	 */
	for (i = 0; i < bufmap->page_count; i++)
		flush_dcache_page(bufmap->page_array[i]);

	/* build a list of available descriptors */
	for (offset = 0, i = 0; i < bufmap->desc_count; i++) {
		bufmap->desc_array[i].page_array = &bufmap->page_array[offset];
		bufmap->desc_array[i].array_count = pages_per_desc;
		bufmap->desc_array[i].uaddr =
		    (user_desc->ptr + (i * pages_per_desc * PAGE_SIZE));
		offset += pages_per_desc;
	}

	return 0;
}

/*
 * orangefs_bufmap_initialize()
 *
 * initializes the mapped buffer interface
 *
 * returns 0 on success, -errno on failure
 */
int orangefs_bufmap_initialize(struct ORANGEFS_dev_map_desc *user_desc)
{
	struct orangefs_bufmap *bufmap;
	int ret = -EINVAL;

	gossip_debug(GOSSIP_BUFMAP_DEBUG,
		     "orangefs_bufmap_initialize: called (ptr ("
		     "%p) sz (%d) cnt(%d).\n",
		     user_desc->ptr,
		     user_desc->size,
		     user_desc->count);

	if (user_desc->total_size < 0 ||
	    user_desc->size < 0 ||
	    user_desc->count < 0)
		goto out;

	/*
	 * sanity check alignment and size of buffer that caller wants to
	 * work with
	 */
	if (PAGE_ALIGN((unsigned long)user_desc->ptr) !=
	    (unsigned long)user_desc->ptr) {
		gossip_err("orangefs error: memory alignment (front). %p\n",
			   user_desc->ptr);
		goto out;
	}

	if (PAGE_ALIGN(((unsigned long)user_desc->ptr + user_desc->total_size))
	    != (unsigned long)(user_desc->ptr + user_desc->total_size)) {
		gossip_err("orangefs error: memory alignment (back).(%p + %d)\n",
			   user_desc->ptr,
			   user_desc->total_size);
		goto out;
	}

	if (user_desc->total_size != (user_desc->size * user_desc->count)) {
		gossip_err("orangefs error: user provided an oddly sized buffer: (%d, %d, %d)\n",
			   user_desc->total_size,
			   user_desc->size,
			   user_desc->count);
		goto out;
	}

	if ((user_desc->size % PAGE_SIZE) != 0) {
		gossip_err("orangefs error: bufmap size not page size divisible (%d).\n",
			   user_desc->size);
		goto out;
	}

	ret = -ENOMEM;
	bufmap = orangefs_bufmap_alloc(user_desc);
	if (!bufmap)
		goto out;

	ret = orangefs_bufmap_map(bufmap, user_desc);
	if (ret)
		goto out_free_bufmap;


	spin_lock(&orangefs_bufmap_lock);
	if (__orangefs_bufmap) {
		spin_unlock(&orangefs_bufmap_lock);
		gossip_err("orangefs: error: bufmap already initialized.\n");
		ret = -EINVAL;
		goto out_unmap_bufmap;
	}
	__orangefs_bufmap = bufmap;
	install(&rw_map,
		bufmap->desc_count,
		bufmap->buffer_index_array);
	install(&readdir_map,
		ORANGEFS_READDIR_DEFAULT_DESC_COUNT,
		bufmap->readdir_index_array);
	spin_unlock(&orangefs_bufmap_lock);

	gossip_debug(GOSSIP_BUFMAP_DEBUG,
		     "orangefs_bufmap_initialize: exiting normally\n");
	return 0;

out_unmap_bufmap:
	orangefs_bufmap_unmap(bufmap);
out_free_bufmap:
	orangefs_bufmap_free(bufmap);
out:
	return ret;
}

/*
 * orangefs_bufmap_finalize()
 *
 * shuts down the mapped buffer interface and releases any resources
 * associated with it
 *
 * no return value
 */
void orangefs_bufmap_finalize(void)
{
	struct orangefs_bufmap *bufmap = __orangefs_bufmap;
	if (!bufmap)
		return;
	gossip_debug(GOSSIP_BUFMAP_DEBUG, "orangefs_bufmap_finalize: called\n");
	mark_killed(&rw_map);
	mark_killed(&readdir_map);
	gossip_debug(GOSSIP_BUFMAP_DEBUG,
		     "orangefs_bufmap_finalize: exiting normally\n");
}

void orangefs_bufmap_run_down(void)
{
	struct orangefs_bufmap *bufmap = __orangefs_bufmap;
	if (!bufmap)
		return;
	run_down(&rw_map);
	run_down(&readdir_map);
	spin_lock(&orangefs_bufmap_lock);
	__orangefs_bufmap = NULL;
	spin_unlock(&orangefs_bufmap_lock);
	orangefs_bufmap_unmap(bufmap);
	orangefs_bufmap_free(bufmap);
}

/*
 * orangefs_bufmap_get()
 *
 * gets a free mapped buffer descriptor, will sleep until one becomes
 * available if necessary
 *
 * returns slot on success, -errno on failure
 */
int orangefs_bufmap_get(void)
{
	return get(&rw_map);
}

/*
 * orangefs_bufmap_put()
 *
 * returns a mapped buffer descriptor to the collection
 *
 * no return value
 */
void orangefs_bufmap_put(int buffer_index)
{
	put(&rw_map, buffer_index);
}

/*
 * orangefs_readdir_index_get()
 *
 * gets a free descriptor, will sleep until one becomes
 * available if necessary.
 * Although the readdir buffers are not mapped into kernel space
 * we could do that at a later point of time. Regardless, these
 * indices are used by the client-core.
 *
 * returns slot on success, -errno on failure
 */
int orangefs_readdir_index_get(void)
{
	return get(&readdir_map);
}

void orangefs_readdir_index_put(int buffer_index)
{
	put(&readdir_map, buffer_index);
}

/*
 * we've been handed an iovec, we need to copy it to 
 * the shared memory descriptor at "buffer_index".
 */
int orangefs_bufmap_copy_from_iovec(struct iov_iter *iter,
				int buffer_index,
				size_t size)
{
	struct orangefs_bufmap_desc *to;
	int i;

	gossip_debug(GOSSIP_BUFMAP_DEBUG,
		     "%s: buffer_index:%d: size:%zu:\n",
		     __func__, buffer_index, size);

	to = &__orangefs_bufmap->desc_array[buffer_index];
	for (i = 0; size; i++) {
		struct page *page = to->page_array[i];
		size_t n = size;
		if (n > PAGE_SIZE)
			n = PAGE_SIZE;
		if (copy_page_from_iter(page, 0, n, iter) != n)
			return -EFAULT;
		size -= n;
	}
	return 0;
}

/*
 * we've been handed an iovec, we need to fill it from
 * the shared memory descriptor at "buffer_index".
 */
int orangefs_bufmap_copy_to_iovec(struct iov_iter *iter,
				    int buffer_index,
				    size_t size)
{
	struct orangefs_bufmap_desc *from;
	int i;

	from = &__orangefs_bufmap->desc_array[buffer_index];
	gossip_debug(GOSSIP_BUFMAP_DEBUG,
		     "%s: buffer_index:%d: size:%zu:\n",
		     __func__, buffer_index, size);


	for (i = 0; size; i++) {
		struct page *page = from->page_array[i];
		size_t n = size;
		if (n > PAGE_SIZE)
			n = PAGE_SIZE;
		n = copy_page_to_iter(page, 0, n, iter);
		if (!n)
			return -EFAULT;
		size -= n;
	}
	return 0;
}
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
274dcf55 MM	2	/*
	3	* (C) 2001 Clemson University and The University of Chicago
	4	*
	5	* See COPYING in top-level directory.
	6	*/
	7	#include "protocol.h"
575e9461 MM	8	#include "orangefs-kernel.h"
575e9461 MM	9	#include "orangefs-bufmap.h"
274dcf55	10
ea2c9c9f AV	11	struct slot_map {
	12	int c;
	13	wait_queue_head_t q;
	14	int count;
	15	unsigned long *map;
	16	};
	17
	18	static struct slot_map rw_map = {
	19	.c = -1,
	20	.q = __WAIT_QUEUE_HEAD_INITIALIZER(rw_map.q)
	21	};
	22	static struct slot_map readdir_map = {
	23	.c = -1,
	24	.q = __WAIT_QUEUE_HEAD_INITIALIZER(readdir_map.q)
	25	};
	26
	27
	28	static void install(struct slot_map m, int count, unsigned long map)
	29	{
	30	spin_lock(&m->q.lock);
	31	m->c = m->count = count;
	32	m->map = map;
	33	wake_up_all_locked(&m->q);
	34	spin_unlock(&m->q.lock);
	35	}
	36
	37	static void mark_killed(struct slot_map *m)
	38	{
	39	spin_lock(&m->q.lock);
	40	m->c -= m->count + 1;
	41	spin_unlock(&m->q.lock);
	42	}
	43
	44	static void run_down(struct slot_map *m)
	45	{
	46	DEFINE_WAIT(wait);
	47	spin_lock(&m->q.lock);
	48	if (m->c != -1) {
	49	for (;;) {
2055da97	50	if (likely(list_empty(&wait.entry)))
ac6424b9	51	__add_wait_queue_entry_tail(&m->q, &wait);
ea2c9c9f AV	52	set_current_state(TASK_UNINTERRUPTIBLE);
	53
	54	if (m->c == -1)
	55	break;
	56
	57	spin_unlock(&m->q.lock);
	58	schedule();
	59	spin_lock(&m->q.lock);
	60	}
	61	__remove_wait_queue(&m->q, &wait);
	62	__set_current_state(TASK_RUNNING);
	63	}
	64	m->map = NULL;
	65	spin_unlock(&m->q.lock);
	66	}
	67
	68	static void put(struct slot_map *m, int slot)
	69	{
	70	int v;
	71	spin_lock(&m->q.lock);
	72	__clear_bit(slot, m->map);
	73	v = ++m->c;
	74	if (unlikely(v == 1)) /* no free slots -> one free slot */
	75	wake_up_locked(&m->q);
	76	else if (unlikely(v == -1)) /* finished dying */
	77	wake_up_all_locked(&m->q);
	78	spin_unlock(&m->q.lock);
	79	}
	80
	81	static int wait_for_free(struct slot_map *m)
	82	{
	83	long left = slot_timeout_secs * HZ;
	84	DEFINE_WAIT(wait);
	85
	86	do {
	87	long n = left, t;
2055da97	88	if (likely(list_empty(&wait.entry)))
ac6424b9	89	__add_wait_queue_entry_tail_exclusive(&m->q, &wait);
ea2c9c9f AV	90	set_current_state(TASK_INTERRUPTIBLE);
	91
	92	if (m->c > 0)
	93	break;
	94
	95	if (m->c < 0) {
	96	/* we are waiting for map to be installed */
	97	/* it would better be there soon, or we go away */
	98	if (n > ORANGEFS_BUFMAP_WAIT_TIMEOUT_SECS * HZ)
	99	n = ORANGEFS_BUFMAP_WAIT_TIMEOUT_SECS * HZ;
	100	}
	101	spin_unlock(&m->q.lock);
	102	t = schedule_timeout(n);
	103	spin_lock(&m->q.lock);
	104	if (unlikely(!t) && n != left && m->c < 0)
	105	left = t;
	106	else
	107	left = t + (left - n);
	108	if (unlikely(signal_pending(current)))
	109	left = -EINTR;
	110	} while (left > 0);
	111
2055da97 IM	112	if (!list_empty(&wait.entry))
2055da97 IM	113	list_del(&wait.entry);
ea2c9c9f AV	114	else if (left <= 0 && waitqueue_active(&m->q))
	115	__wake_up_locked_key(&m->q, TASK_INTERRUPTIBLE, NULL);
	116	__set_current_state(TASK_RUNNING);
	117
	118	if (likely(left > 0))
	119	return 0;
	120
	121	return left < 0 ? -EINTR : -ETIMEDOUT;
	122	}
	123
	124	static int get(struct slot_map *m)
	125	{
	126	int res = 0;
	127	spin_lock(&m->q.lock);
	128	if (unlikely(m->c <= 0))
	129	res = wait_for_free(m);
	130	if (likely(!res)) {
	131	m->c--;
	132	res = find_first_zero_bit(m->map, m->count);
	133	__set_bit(res, m->map);
	134	}
	135	spin_unlock(&m->q.lock);
	136	return res;
	137	}
274dcf55	138
bf89f584 MB	139	/* used to describe mapped buffers */
	140	struct orangefs_bufmap_desc {
	141	void uaddr; / user space address pointer */
	142	struct page *page_array; / array of mapped pages */
	143	int array_count; /* size of above arrays */
	144	struct list_head list_link;
	145	};
	146
8bb8aefd	147	static struct orangefs_bufmap {
274dcf55 MM	148	int desc_size;
	149	int desc_shift;
	150	int desc_count;
	151	int total_size;
	152	int page_count;
	153
	154	struct page **page_array;
8bb8aefd	155	struct orangefs_bufmap_desc *desc_array;
274dcf55 MM	156
274dcf55 MM	157	/* array to track usage of buffer descriptors */
ea2c9c9f	158	unsigned long *buffer_index_array;
274dcf55 MM	159
274dcf55 MM	160	/* array to track usage of buffer descriptors for readdir */
ea2c9c9f AV	161	#define N DIV_ROUND_UP(ORANGEFS_READDIR_DEFAULT_DESC_COUNT, BITS_PER_LONG)
	162	unsigned long readdir_index_array[N];
	163	#undef N
8bb8aefd	164	} *__orangefs_bufmap;
274dcf55	165
8bb8aefd	166	static DEFINE_SPINLOCK(orangefs_bufmap_lock);
274dcf55 MM	167
274dcf55 MM	168	static void
8bb8aefd	169	orangefs_bufmap_unmap(struct orangefs_bufmap *bufmap)
274dcf55 MM	170	{
	171	int i;
	172
	173	for (i = 0; i < bufmap->page_count; i++)
09cbfeaf	174	put_page(bufmap->page_array[i]);
274dcf55 MM	175	}
	176
	177	static void
8bb8aefd	178	orangefs_bufmap_free(struct orangefs_bufmap *bufmap)
274dcf55 MM	179	{
	180	kfree(bufmap->page_array);
	181	kfree(bufmap->desc_array);
	182	kfree(bufmap->buffer_index_array);
	183	kfree(bufmap);
	184	}
	185
b09d10df MB	186	/*
	187	* XXX: Can the size and shift change while the caller gives up the
	188	* XXX: lock between calling this and doing something useful?
	189	*/
	190
765a75b3	191	int orangefs_bufmap_size_query(void)
274dcf55	192	{
b09d10df MB	193	struct orangefs_bufmap *bufmap;
b09d10df MB	194	int size = 0;
17804184 AV	195	spin_lock(&orangefs_bufmap_lock);
	196	bufmap = __orangefs_bufmap;
	197	if (bufmap)
b09d10df	198	size = bufmap->desc_size;
17804184	199	spin_unlock(&orangefs_bufmap_lock);
274dcf55 MM	200	return size;
	201	}
	202
765a75b3	203	int orangefs_bufmap_shift_query(void)
274dcf55	204	{
b09d10df MB	205	struct orangefs_bufmap *bufmap;
b09d10df MB	206	int shift = 0;
17804184 AV	207	spin_lock(&orangefs_bufmap_lock);
	208	bufmap = __orangefs_bufmap;
	209	if (bufmap)
b09d10df	210	shift = bufmap->desc_shift;
17804184	211	spin_unlock(&orangefs_bufmap_lock);
274dcf55 MM	212	return shift;
	213	}
	214
	215	static DECLARE_WAIT_QUEUE_HEAD(bufmap_waitq);
	216	static DECLARE_WAIT_QUEUE_HEAD(readdir_waitq);
	217
	218	/*
7d221485	219	* orangefs_get_bufmap_init
274dcf55 MM	220	*
	221	* If bufmap_init is 1, then the shared memory system, including the
	222	* buffer_index_array, is available. Otherwise, it is not.
	223	*
	224	* returns the value of bufmap_init
	225	*/
7d221485	226	int orangefs_get_bufmap_init(void)
274dcf55	227	{
8bb8aefd	228	return __orangefs_bufmap ? 1 : 0;
274dcf55 MM	229	}
	230
	231
8bb8aefd YL	232	static struct orangefs_bufmap *
8bb8aefd YL	233	orangefs_bufmap_alloc(struct ORANGEFS_dev_map_desc *user_desc)
274dcf55	234	{
8bb8aefd	235	struct orangefs_bufmap *bufmap;
274dcf55 MM	236
	237	bufmap = kzalloc(sizeof(*bufmap), GFP_KERNEL);
	238	if (!bufmap)
	239	goto out;
	240
274dcf55 MM	241	bufmap->total_size = user_desc->total_size;
	242	bufmap->desc_count = user_desc->count;
	243	bufmap->desc_size = user_desc->size;
	244	bufmap->desc_shift = ilog2(bufmap->desc_size);
	245
274dcf55	246	bufmap->buffer_index_array =
ea2c9c9f	247	kzalloc(DIV_ROUND_UP(bufmap->desc_count, BITS_PER_LONG), GFP_KERNEL);
07a25853	248	if (!bufmap->buffer_index_array)
274dcf55	249	goto out_free_bufmap;
274dcf55 MM	250
274dcf55 MM	251	bufmap->desc_array =
8bb8aefd	252	kcalloc(bufmap->desc_count, sizeof(struct orangefs_bufmap_desc),
274dcf55	253	GFP_KERNEL);
07a25853	254	if (!bufmap->desc_array)
274dcf55	255	goto out_free_index_array;
274dcf55 MM	256
	257	bufmap->page_count = bufmap->total_size / PAGE_SIZE;
	258
	259	/* allocate storage to track our page mappings */
	260	bufmap->page_array =
	261	kcalloc(bufmap->page_count, sizeof(struct page *), GFP_KERNEL);
	262	if (!bufmap->page_array)
	263	goto out_free_desc_array;
	264
	265	return bufmap;
	266
	267	out_free_desc_array:
	268	kfree(bufmap->desc_array);
	269	out_free_index_array:
	270	kfree(bufmap->buffer_index_array);
	271	out_free_bufmap:
	272	kfree(bufmap);
	273	out:
	274	return NULL;
	275	}
	276
	277	static int
8bb8aefd YL	278	orangefs_bufmap_map(struct orangefs_bufmap *bufmap,
8bb8aefd YL	279	struct ORANGEFS_dev_map_desc *user_desc)
274dcf55 MM	280	{
	281	int pages_per_desc = bufmap->desc_size / PAGE_SIZE;
	282	int offset = 0, ret, i;
	283
	284	/* map the pages */
16742f2d AV	285	ret = get_user_pages_fast((unsigned long)user_desc->ptr,
16742f2d AV	286	bufmap->page_count, 1, bufmap->page_array);
274dcf55 MM	287
	288	if (ret < 0)
	289	return ret;
	290
	291	if (ret != bufmap->page_count) {
8bb8aefd	292	gossip_err("orangefs error: asked for %d pages, only got %d.\n",
274dcf55 MM	293	bufmap->page_count, ret);
	294
	295	for (i = 0; i < ret; i++) {
	296	SetPageError(bufmap->page_array[i]);
09cbfeaf	297	put_page(bufmap->page_array[i]);
274dcf55 MM	298	}
	299	return -ENOMEM;
	300	}
	301
	302	/*
	303	* ideally we want to get kernel space pointers for each page, but
	304	* we can't kmap that many pages at once if highmem is being used.
	305	* so instead, we just kmap/kunmap the page address each time the
	306	* kaddr is needed.
	307	*/
	308	for (i = 0; i < bufmap->page_count; i++)
	309	flush_dcache_page(bufmap->page_array[i]);
	310
	311	/* build a list of available descriptors */
	312	for (offset = 0, i = 0; i < bufmap->desc_count; i++) {
	313	bufmap->desc_array[i].page_array = &bufmap->page_array[offset];
	314	bufmap->desc_array[i].array_count = pages_per_desc;
	315	bufmap->desc_array[i].uaddr =
	316	(user_desc->ptr + (i * pages_per_desc * PAGE_SIZE));
	317	offset += pages_per_desc;
	318	}
	319
	320	return 0;
	321	}
	322
	323	/*
8bb8aefd	324	* orangefs_bufmap_initialize()
274dcf55 MM	325	*
	326	* initializes the mapped buffer interface
	327	*
	328	* returns 0 on success, -errno on failure
	329	*/
8bb8aefd	330	int orangefs_bufmap_initialize(struct ORANGEFS_dev_map_desc *user_desc)
274dcf55	331	{
8bb8aefd	332	struct orangefs_bufmap *bufmap;
274dcf55 MM	333	int ret = -EINVAL;
	334
	335	gossip_debug(GOSSIP_BUFMAP_DEBUG,
8bb8aefd	336	"orangefs_bufmap_initialize: called (ptr ("
274dcf55 MM	337	"%p) sz (%d) cnt(%d).\n",
	338	user_desc->ptr,
	339	user_desc->size,
	340	user_desc->count);
	341
eb82fbcf DC	342	if (user_desc->total_size < 0 \|\|
	343	user_desc->size < 0 \|\|
	344	user_desc->count < 0)
	345	goto out;
	346
274dcf55 MM	347	/*
	348	* sanity check alignment and size of buffer that caller wants to
	349	* work with
	350	*/
	351	if (PAGE_ALIGN((unsigned long)user_desc->ptr) !=
	352	(unsigned long)user_desc->ptr) {
8bb8aefd	353	gossip_err("orangefs error: memory alignment (front). %p\n",
274dcf55 MM	354	user_desc->ptr);
	355	goto out;
	356	}
	357
	358	if (PAGE_ALIGN(((unsigned long)user_desc->ptr + user_desc->total_size))
	359	!= (unsigned long)(user_desc->ptr + user_desc->total_size)) {
8bb8aefd	360	gossip_err("orangefs error: memory alignment (back).(%p + %d)\n",
274dcf55 MM	361	user_desc->ptr,
	362	user_desc->total_size);
	363	goto out;
	364	}
	365
	366	if (user_desc->total_size != (user_desc->size * user_desc->count)) {
8bb8aefd	367	gossip_err("orangefs error: user provided an oddly sized buffer: (%d, %d, %d)\n",
274dcf55 MM	368	user_desc->total_size,
	369	user_desc->size,
	370	user_desc->count);
	371	goto out;
	372	}
	373
	374	if ((user_desc->size % PAGE_SIZE) != 0) {
8bb8aefd	375	gossip_err("orangefs error: bufmap size not page size divisible (%d).\n",
274dcf55 MM	376	user_desc->size);
	377	goto out;
	378	}
	379
	380	ret = -ENOMEM;
8bb8aefd	381	bufmap = orangefs_bufmap_alloc(user_desc);
274dcf55 MM	382	if (!bufmap)
	383	goto out;
	384
8bb8aefd	385	ret = orangefs_bufmap_map(bufmap, user_desc);
274dcf55 MM	386	if (ret)
	387	goto out_free_bufmap;
	388
	389
8bb8aefd YL	390	spin_lock(&orangefs_bufmap_lock);
	391	if (__orangefs_bufmap) {
	392	spin_unlock(&orangefs_bufmap_lock);
	393	gossip_err("orangefs: error: bufmap already initialized.\n");
ea2c9c9f	394	ret = -EINVAL;
274dcf55 MM	395	goto out_unmap_bufmap;
274dcf55 MM	396	}
8bb8aefd	397	__orangefs_bufmap = bufmap;
ea2c9c9f AV	398	install(&rw_map,
	399	bufmap->desc_count,
	400	bufmap->buffer_index_array);
	401	install(&readdir_map,
	402	ORANGEFS_READDIR_DEFAULT_DESC_COUNT,
	403	bufmap->readdir_index_array);
8bb8aefd	404	spin_unlock(&orangefs_bufmap_lock);
274dcf55	405
274dcf55	406	gossip_debug(GOSSIP_BUFMAP_DEBUG,
8bb8aefd	407	"orangefs_bufmap_initialize: exiting normally\n");
274dcf55 MM	408	return 0;
	409
	410	out_unmap_bufmap:
8bb8aefd	411	orangefs_bufmap_unmap(bufmap);
274dcf55	412	out_free_bufmap:
8bb8aefd	413	orangefs_bufmap_free(bufmap);
274dcf55 MM	414	out:
	415	return ret;
	416	}
	417
	418	/*
8bb8aefd	419	* orangefs_bufmap_finalize()
274dcf55 MM	420	*
	421	* shuts down the mapped buffer interface and releases any resources
	422	* associated with it
	423	*
	424	* no return value
	425	*/
8bb8aefd	426	void orangefs_bufmap_finalize(void)
274dcf55	427	{
ea2c9c9f AV	428	struct orangefs_bufmap *bufmap = __orangefs_bufmap;
	429	if (!bufmap)
	430	return;
8bb8aefd	431	gossip_debug(GOSSIP_BUFMAP_DEBUG, "orangefs_bufmap_finalize: called\n");
ea2c9c9f AV	432	mark_killed(&rw_map);
ea2c9c9f AV	433	mark_killed(&readdir_map);
274dcf55	434	gossip_debug(GOSSIP_BUFMAP_DEBUG,
8bb8aefd	435	"orangefs_bufmap_finalize: exiting normally\n");
274dcf55 MM	436	}
274dcf55 MM	437
ea2c9c9f	438	void orangefs_bufmap_run_down(void)
274dcf55	439	{
ea2c9c9f AV	440	struct orangefs_bufmap *bufmap = __orangefs_bufmap;
ea2c9c9f AV	441	if (!bufmap)
274dcf55	442	return;
ea2c9c9f AV	443	run_down(&rw_map);
	444	run_down(&readdir_map);
	445	spin_lock(&orangefs_bufmap_lock);
	446	__orangefs_bufmap = NULL;
	447	spin_unlock(&orangefs_bufmap_lock);
	448	orangefs_bufmap_unmap(bufmap);
	449	orangefs_bufmap_free(bufmap);
274dcf55 MM	450	}
	451
	452	/*
8bb8aefd	453	* orangefs_bufmap_get()
274dcf55 MM	454	*
	455	* gets a free mapped buffer descriptor, will sleep until one becomes
	456	* available if necessary
	457	*
b8a99a8f	458	* returns slot on success, -errno on failure
274dcf55	459	*/
b8a99a8f	460	int orangefs_bufmap_get(void)
274dcf55	461	{
b8a99a8f	462	return get(&rw_map);
274dcf55 MM	463	}
	464
	465	/*
8bb8aefd	466	* orangefs_bufmap_put()
274dcf55 MM	467	*
	468	* returns a mapped buffer descriptor to the collection
	469	*
	470	* no return value
	471	*/
1357d06d	472	void orangefs_bufmap_put(int buffer_index)
274dcf55	473	{
ea2c9c9f	474	put(&rw_map, buffer_index);
274dcf55 MM	475	}
	476
	477	/*
7d221485	478	* orangefs_readdir_index_get()
274dcf55 MM	479	*
	480	* gets a free descriptor, will sleep until one becomes
	481	* available if necessary.
	482	* Although the readdir buffers are not mapped into kernel space
	483	* we could do that at a later point of time. Regardless, these
	484	* indices are used by the client-core.
	485	*
b8a99a8f	486	* returns slot on success, -errno on failure
274dcf55	487	*/
b8a99a8f	488	int orangefs_readdir_index_get(void)
274dcf55	489	{
b8a99a8f	490	return get(&readdir_map);
274dcf55 MM	491	}
274dcf55 MM	492
82d37f19	493	void orangefs_readdir_index_put(int buffer_index)
274dcf55	494	{
ea2c9c9f	495	put(&readdir_map, buffer_index);
274dcf55 MM	496	}
274dcf55 MM	497
b5e376ea MM	498	/*
	499	* we've been handed an iovec, we need to copy it to
	500	* the shared memory descriptor at "buffer_index".
	501	*/
bf6bf606	502	int orangefs_bufmap_copy_from_iovec(struct iov_iter *iter,
54804949 MM	503	int buffer_index,
54804949 MM	504	size_t size)
274dcf55	505	{
bf6bf606	506	struct orangefs_bufmap_desc *to;
4d1c4404	507	int i;
274dcf55 MM	508
274dcf55 MM	509	gossip_debug(GOSSIP_BUFMAP_DEBUG,
34204fde	510	"%s: buffer_index:%d: size:%zu:\n",
4d1c4404	511	__func__, buffer_index, size);
274dcf55	512
bf6bf606	513	to = &__orangefs_bufmap->desc_array[buffer_index];
4d1c4404	514	for (i = 0; size; i++) {
34204fde AV	515	struct page *page = to->page_array[i];
	516	size_t n = size;
	517	if (n > PAGE_SIZE)
	518	n = PAGE_SIZE;
890559e3	519	if (copy_page_from_iter(page, 0, n, iter) != n)
34204fde AV	520	return -EFAULT;
34204fde AV	521	size -= n;
274dcf55	522	}
34204fde	523	return 0;
274dcf55 MM	524	}
	525
	526	/*
b5e376ea MM	527	* we've been handed an iovec, we need to fill it from
b5e376ea MM	528	* the shared memory descriptor at "buffer_index".
274dcf55	529	*/
bf6bf606	530	int orangefs_bufmap_copy_to_iovec(struct iov_iter *iter,
5c278228 AV	531	int buffer_index,
5c278228 AV	532	size_t size)
274dcf55	533	{
bf6bf606	534	struct orangefs_bufmap_desc *from;
4d1c4404	535	int i;
274dcf55	536
bf6bf606	537	from = &__orangefs_bufmap->desc_array[buffer_index];
274dcf55	538	gossip_debug(GOSSIP_BUFMAP_DEBUG,
5c278228 AV	539	"%s: buffer_index:%d: size:%zu:\n",
5c278228 AV	540	__func__, buffer_index, size);
274dcf55	541
274dcf55	542
5c278228 AV	543	for (i = 0; size; i++) {
	544	struct page *page = from->page_array[i];
	545	size_t n = size;
	546	if (n > PAGE_SIZE)
	547	n = PAGE_SIZE;
	548	n = copy_page_to_iter(page, 0, n, iter);
	549	if (!n)
	550	return -EFAULT;
	551	size -= n;
274dcf55	552	}
5c278228	553	return 0;
274dcf55	554	}