[people/arne_f/kernel.git] / mm / iov_iter.c

#include <linux/export.h>
#include <linux/uio.h>
#include <linux/pagemap.h>

size_t copy_page_to_iter(struct page *page, size_t offset, size_t bytes,
			 struct iov_iter *i)
{
	size_t skip, copy, left, wanted;
	const struct iovec *iov;
	char __user *buf;
	void *kaddr, *from;

	if (unlikely(bytes > i->count))
		bytes = i->count;

	if (unlikely(!bytes))
		return 0;

	wanted = bytes;
	iov = i->iov;
	skip = i->iov_offset;
	buf = iov->iov_base + skip;
	copy = min(bytes, iov->iov_len - skip);

	if (!fault_in_pages_writeable(buf, copy)) {
		kaddr = kmap_atomic(page);
		from = kaddr + offset;

		/* first chunk, usually the only one */
		left = __copy_to_user_inatomic(buf, from, copy);
		copy -= left;
		skip += copy;
		from += copy;
		bytes -= copy;

		while (unlikely(!left && bytes)) {
			iov++;
			buf = iov->iov_base;
			copy = min(bytes, iov->iov_len);
			left = __copy_to_user_inatomic(buf, from, copy);
			copy -= left;
			skip = copy;
			from += copy;
			bytes -= copy;
		}
		if (likely(!bytes)) {
			kunmap_atomic(kaddr);
			goto done;
		}
		offset = from - kaddr;
		buf += copy;
		kunmap_atomic(kaddr);
		copy = min(bytes, iov->iov_len - skip);
	}
	/* Too bad - revert to non-atomic kmap */
	kaddr = kmap(page);
	from = kaddr + offset;
	left = __copy_to_user(buf, from, copy);
	copy -= left;
	skip += copy;
	from += copy;
	bytes -= copy;
	while (unlikely(!left && bytes)) {
		iov++;
		buf = iov->iov_base;
		copy = min(bytes, iov->iov_len);
		left = __copy_to_user(buf, from, copy);
		copy -= left;
		skip = copy;
		from += copy;
		bytes -= copy;
	}
	kunmap(page);
done:
	i->count -= wanted - bytes;
	i->nr_segs -= iov - i->iov;
	i->iov = iov;
	i->iov_offset = skip;
	return wanted - bytes;
}
EXPORT_SYMBOL(copy_page_to_iter);

static size_t __iovec_copy_from_user_inatomic(char *vaddr,
			const struct iovec *iov, size_t base, size_t bytes)
{
	size_t copied = 0, left = 0;

	while (bytes) {
		char __user *buf = iov->iov_base + base;
		int copy = min(bytes, iov->iov_len - base);

		base = 0;
		left = __copy_from_user_inatomic(vaddr, buf, copy);
		copied += copy;
		bytes -= copy;
		vaddr += copy;
		iov++;

		if (unlikely(left))
			break;
	}
	return copied - left;
}

/*
 * Copy as much as we can into the page and return the number of bytes which
 * were successfully copied.  If a fault is encountered then return the number of
 * bytes which were copied.
 */
size_t iov_iter_copy_from_user_atomic(struct page *page,
		struct iov_iter *i, unsigned long offset, size_t bytes)
{
	char *kaddr;
	size_t copied;

	kaddr = kmap_atomic(page);
	if (likely(i->nr_segs == 1)) {
		int left;
		char __user *buf = i->iov->iov_base + i->iov_offset;
		left = __copy_from_user_inatomic(kaddr + offset, buf, bytes);
		copied = bytes - left;
	} else {
		copied = __iovec_copy_from_user_inatomic(kaddr + offset,
						i->iov, i->iov_offset, bytes);
	}
	kunmap_atomic(kaddr);

	return copied;
}
EXPORT_SYMBOL(iov_iter_copy_from_user_atomic);

void iov_iter_advance(struct iov_iter *i, size_t bytes)
{
	BUG_ON(i->count < bytes);

	if (likely(i->nr_segs == 1)) {
		i->iov_offset += bytes;
		i->count -= bytes;
	} else {
		const struct iovec *iov = i->iov;
		size_t base = i->iov_offset;
		unsigned long nr_segs = i->nr_segs;

		/*
		 * The !iov->iov_len check ensures we skip over unlikely
		 * zero-length segments (without overruning the iovec).
		 */
		while (bytes || unlikely(i->count && !iov->iov_len)) {
			int copy;

			copy = min(bytes, iov->iov_len - base);
			BUG_ON(!i->count || i->count < copy);
			i->count -= copy;
			bytes -= copy;
			base += copy;
			if (iov->iov_len == base) {
				iov++;
				nr_segs--;
				base = 0;
			}
		}
		i->iov = iov;
		i->iov_offset = base;
		i->nr_segs = nr_segs;
	}
}
EXPORT_SYMBOL(iov_iter_advance);

/*
 * Fault in the first iovec of the given iov_iter, to a maximum length
 * of bytes. Returns 0 on success, or non-zero if the memory could not be
 * accessed (ie. because it is an invalid address).
 *
 * writev-intensive code may want this to prefault several iovecs -- that
 * would be possible (callers must not rely on the fact that _only_ the
 * first iovec will be faulted with the current implementation).
 */
int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes)
{
	char __user *buf = i->iov->iov_base + i->iov_offset;
	bytes = min(bytes, i->iov->iov_len - i->iov_offset);
	return fault_in_pages_readable(buf, bytes);
}
EXPORT_SYMBOL(iov_iter_fault_in_readable);

/*
 * Return the count of just the current iov_iter segment.
 */
size_t iov_iter_single_seg_count(const struct iov_iter *i)
{
	const struct iovec *iov = i->iov;
	if (i->nr_segs == 1)
		return i->count;
	else
		return min(i->count, iov->iov_len - i->iov_offset);
}
EXPORT_SYMBOL(iov_iter_single_seg_count);

unsigned long iov_iter_alignment(const struct iov_iter *i)
{
	const struct iovec *iov = i->iov;
	unsigned long res;
	size_t size = i->count;
	size_t n;

	if (!size)
		return 0;

	res = (unsigned long)iov->iov_base + i->iov_offset;
	n = iov->iov_len - i->iov_offset;
	if (n >= size)
		return res | size;
	size -= n;
	res |= n;
	while (size > (++iov)->iov_len) {
		res |= (unsigned long)iov->iov_base | iov->iov_len;
		size -= iov->iov_len;
	}
	res |= (unsigned long)iov->iov_base | size;
	return res;
}
EXPORT_SYMBOL(iov_iter_alignment);

void iov_iter_init(struct iov_iter *i, int direction,
			const struct iovec *iov, unsigned long nr_segs,
			size_t count)
{
	/* It will get better.  Eventually... */
	if (segment_eq(get_fs(), KERNEL_DS))
		direction |= REQ_KERNEL;
	i->type = direction;
	i->iov = iov;
	i->nr_segs = nr_segs;
	i->iov_offset = 0;
	i->count = count;
}
EXPORT_SYMBOL(iov_iter_init);

ssize_t iov_iter_get_pages(struct iov_iter *i,
		   struct page **pages, size_t maxsize,
		   size_t *start)
{
	size_t offset = i->iov_offset;
	const struct iovec *iov = i->iov;
	size_t len;
	unsigned long addr;
	int n;
	int res;

	len = iov->iov_len - offset;
	if (len > i->count)
		len = i->count;
	if (len > maxsize)
		len = maxsize;
	addr = (unsigned long)iov->iov_base + offset;
	len += *start = addr & (PAGE_SIZE - 1);
	addr &= ~(PAGE_SIZE - 1);
	n = (len + PAGE_SIZE - 1) / PAGE_SIZE;
	res = get_user_pages_fast(addr, n, (i->type & WRITE) != WRITE, pages);
	if (unlikely(res < 0))
		return res;
	return (res == n ? len : res * PAGE_SIZE) - *start;
}
EXPORT_SYMBOL(iov_iter_get_pages);

int iov_iter_npages(const struct iov_iter *i, int maxpages)
{
	size_t offset = i->iov_offset;
	size_t size = i->count;
	const struct iovec *iov = i->iov;
	int npages = 0;
	int n;

	for (n = 0; size && n < i->nr_segs; n++, iov++) {
		unsigned long addr = (unsigned long)iov->iov_base + offset;
		size_t len = iov->iov_len - offset;
		offset = 0;
		if (unlikely(!len))	/* empty segment */
			continue;
		if (len > size)
			len = size;
		npages += (addr + len + PAGE_SIZE - 1) / PAGE_SIZE
			  - addr / PAGE_SIZE;
		if (npages >= maxpages)	/* don't bother going further */
			return maxpages;
		size -= len;
		offset = 0;
	}
	return min(npages, maxpages);
}
EXPORT_SYMBOL(iov_iter_npages);
Commit	Line	Data
4f18cd31 AV	1	#include <linux/export.h>
	2	#include <linux/uio.h>
	3	#include <linux/pagemap.h>
	4
	5	size_t copy_page_to_iter(struct page *page, size_t offset, size_t bytes,
	6	struct iov_iter *i)
	7	{
	8	size_t skip, copy, left, wanted;
	9	const struct iovec *iov;
	10	char __user *buf;
	11	void kaddr, from;
	12
	13	if (unlikely(bytes > i->count))
	14	bytes = i->count;
	15
	16	if (unlikely(!bytes))
	17	return 0;
	18
	19	wanted = bytes;
	20	iov = i->iov;
	21	skip = i->iov_offset;
	22	buf = iov->iov_base + skip;
	23	copy = min(bytes, iov->iov_len - skip);
	24
	25	if (!fault_in_pages_writeable(buf, copy)) {
	26	kaddr = kmap_atomic(page);
	27	from = kaddr + offset;
	28
	29	/* first chunk, usually the only one */
	30	left = __copy_to_user_inatomic(buf, from, copy);
	31	copy -= left;
	32	skip += copy;
	33	from += copy;
	34	bytes -= copy;
	35
	36	while (unlikely(!left && bytes)) {
	37	iov++;
	38	buf = iov->iov_base;
	39	copy = min(bytes, iov->iov_len);
	40	left = __copy_to_user_inatomic(buf, from, copy);
	41	copy -= left;
	42	skip = copy;
	43	from += copy;
	44	bytes -= copy;
	45	}
	46	if (likely(!bytes)) {
	47	kunmap_atomic(kaddr);
	48	goto done;
	49	}
	50	offset = from - kaddr;
	51	buf += copy;
	52	kunmap_atomic(kaddr);
	53	copy = min(bytes, iov->iov_len - skip);
	54	}
	55	/* Too bad - revert to non-atomic kmap */
	56	kaddr = kmap(page);
	57	from = kaddr + offset;
	58	left = __copy_to_user(buf, from, copy);
	59	copy -= left;
	60	skip += copy;
	61	from += copy;
	62	bytes -= copy;
	63	while (unlikely(!left && bytes)) {
	64	iov++;
65	buf = iov->iov_base;
66	copy = min(bytes, iov->iov_len);
67	left = __copy_to_user(buf, from, copy);
68	copy -= left;
69	skip = copy;
70	from += copy;
71	bytes -= copy;
72	}
73	kunmap(page);
74	done:
75	i->count -= wanted - bytes;
76	i->nr_segs -= iov - i->iov;
77	i->iov = iov;
78	i->iov_offset = skip;
79	return wanted - bytes;
80	}
81	EXPORT_SYMBOL(copy_page_to_iter);
82
83	static size_t __iovec_copy_from_user_inatomic(char *vaddr,
84	const struct iovec *iov, size_t base, size_t bytes)
85	{
86	size_t copied = 0, left = 0;
87
88	while (bytes) {
89	char __user *buf = iov->iov_base + base;
90	int copy = min(bytes, iov->iov_len - base);
91
92	base = 0;
93	left = __copy_from_user_inatomic(vaddr, buf, copy);
94	copied += copy;
95	bytes -= copy;
96	vaddr += copy;
97	iov++;
98
99	if (unlikely(left))
100	break;
101	}
102	return copied - left;
103	}
104
105	/*
106	* Copy as much as we can into the page and return the number of bytes which
107	* were successfully copied. If a fault is encountered then return the number of
108	* bytes which were copied.
109	*/
110	size_t iov_iter_copy_from_user_atomic(struct page *page,
111	struct iov_iter *i, unsigned long offset, size_t bytes)
112	{
113	char *kaddr;
114	size_t copied;
115
116	kaddr = kmap_atomic(page);
117	if (likely(i->nr_segs == 1)) {
118	int left;
119	char __user *buf = i->iov->iov_base + i->iov_offset;
120	left = __copy_from_user_inatomic(kaddr + offset, buf, bytes);
121	copied = bytes - left;
122	} else {
123	copied = __iovec_copy_from_user_inatomic(kaddr + offset,
124	i->iov, i->iov_offset, bytes);
125	}
126	kunmap_atomic(kaddr);
127
128	return copied;
129	}
130	EXPORT_SYMBOL(iov_iter_copy_from_user_atomic);
131
4f18cd31 AV	132	void iov_iter_advance(struct iov_iter *i, size_t bytes)
	133	{
	134	BUG_ON(i->count < bytes);
	135
	136	if (likely(i->nr_segs == 1)) {
	137	i->iov_offset += bytes;
	138	i->count -= bytes;
	139	} else {
	140	const struct iovec *iov = i->iov;
	141	size_t base = i->iov_offset;
	142	unsigned long nr_segs = i->nr_segs;
	143
	144	/*
	145	* The !iov->iov_len check ensures we skip over unlikely
	146	* zero-length segments (without overruning the iovec).
	147	*/
	148	while (bytes \|\| unlikely(i->count && !iov->iov_len)) {
	149	int copy;
	150
	151	copy = min(bytes, iov->iov_len - base);
	152	BUG_ON(!i->count \|\| i->count < copy);
	153	i->count -= copy;
	154	bytes -= copy;
	155	base += copy;
	156	if (iov->iov_len == base) {
	157	iov++;
	158	nr_segs--;
	159	base = 0;
	160	}
	161	}
	162	i->iov = iov;
	163	i->iov_offset = base;
	164	i->nr_segs = nr_segs;
	165	}
	166	}
	167	EXPORT_SYMBOL(iov_iter_advance);
	168
	169	/*
	170	* Fault in the first iovec of the given iov_iter, to a maximum length
	171	* of bytes. Returns 0 on success, or non-zero if the memory could not be
	172	* accessed (ie. because it is an invalid address).
	173	*
	174	* writev-intensive code may want this to prefault several iovecs -- that
	175	* would be possible (callers must not rely on the fact that _only_ the
	176	* first iovec will be faulted with the current implementation).
	177	*/
	178	int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes)
	179	{
	180	char __user *buf = i->iov->iov_base + i->iov_offset;
	181	bytes = min(bytes, i->iov->iov_len - i->iov_offset);
	182	return fault_in_pages_readable(buf, bytes);
	183	}
	184	EXPORT_SYMBOL(iov_iter_fault_in_readable);
	185
	186	/*
	187	* Return the count of just the current iov_iter segment.
	188	*/
	189	size_t iov_iter_single_seg_count(const struct iov_iter *i)
	190	{
	191	const struct iovec *iov = i->iov;
	192	if (i->nr_segs == 1)
	193	return i->count;
	194	else
	195	return min(i->count, iov->iov_len - i->iov_offset);
196	}
197	EXPORT_SYMBOL(iov_iter_single_seg_count);
886a3911 AV	198
	199	unsigned long iov_iter_alignment(const struct iov_iter *i)
	200	{
	201	const struct iovec *iov = i->iov;
	202	unsigned long res;
	203	size_t size = i->count;
	204	size_t n;
	205
	206	if (!size)
	207	return 0;
	208
	209	res = (unsigned long)iov->iov_base + i->iov_offset;
	210	n = iov->iov_len - i->iov_offset;
	211	if (n >= size)
	212	return res \| size;
	213	size -= n;
	214	res \|= n;
	215	while (size > (++iov)->iov_len) {
	216	res \|= (unsigned long)iov->iov_base \| iov->iov_len;
	217	size -= iov->iov_len;
	218	}
	219	res \|= (unsigned long)iov->iov_base \| size;
	220	return res;
	221	}
	222	EXPORT_SYMBOL(iov_iter_alignment);
71d8e532 AV	223
	224	void iov_iter_init(struct iov_iter *i, int direction,
	225	const struct iovec *iov, unsigned long nr_segs,
	226	size_t count)
	227	{
	228	/* It will get better. Eventually... */
	229	if (segment_eq(get_fs(), KERNEL_DS))
	230	direction \|= REQ_KERNEL;
	231	i->type = direction;
	232	i->iov = iov;
	233	i->nr_segs = nr_segs;
	234	i->iov_offset = 0;
	235	i->count = count;
	236	}
	237	EXPORT_SYMBOL(iov_iter_init);
7b2c99d1 AV	238
	239	ssize_t iov_iter_get_pages(struct iov_iter *i,
	240	struct page **pages, size_t maxsize,
	241	size_t *start)
	242	{
	243	size_t offset = i->iov_offset;
	244	const struct iovec *iov = i->iov;
	245	size_t len;
	246	unsigned long addr;
	247	int n;
	248	int res;
	249
	250	len = iov->iov_len - offset;
	251	if (len > i->count)
	252	len = i->count;
	253	if (len > maxsize)
	254	len = maxsize;
	255	addr = (unsigned long)iov->iov_base + offset;
	256	len += *start = addr & (PAGE_SIZE - 1);
	257	addr &= ~(PAGE_SIZE - 1);
	258	n = (len + PAGE_SIZE - 1) / PAGE_SIZE;
	259	res = get_user_pages_fast(addr, n, (i->type & WRITE) != WRITE, pages);
	260	if (unlikely(res < 0))
	261	return res;
	262	return (res == n ? len : res * PAGE_SIZE) - *start;
	263	}
	264	EXPORT_SYMBOL(iov_iter_get_pages);
f67da30c AV	265
	266	int iov_iter_npages(const struct iov_iter *i, int maxpages)
	267	{
	268	size_t offset = i->iov_offset;
	269	size_t size = i->count;
	270	const struct iovec *iov = i->iov;
	271	int npages = 0;
	272	int n;
	273
	274	for (n = 0; size && n < i->nr_segs; n++, iov++) {
	275	unsigned long addr = (unsigned long)iov->iov_base + offset;
	276	size_t len = iov->iov_len - offset;
	277	offset = 0;
	278	if (unlikely(!len)) /* empty segment */
	279	continue;
	280	if (len > size)
	281	len = size;
	282	npages += (addr + len + PAGE_SIZE - 1) / PAGE_SIZE
	283	- addr / PAGE_SIZE;
	284	if (npages >= maxpages) /* don't bother going further */
	285	return maxpages;
	286	size -= len;
	287	offset = 0;
	288	}
	289	return min(npages, maxpages);
	290	}
	291	EXPORT_SYMBOL(iov_iter_npages);