new helper: iov_iter_get_pages()

iov_iter_get_pages(iter, pages, maxsize, &start) grabs references pinning
the pages of up to maxsize of (contiguous) data from iter.  Returns the
amount of memory grabbed or -error.  In case of success, the requested
area begins at offset start in pages[0] and runs through pages[1], etc.
Less than requested amount might be returned - either because the contiguous
area in the beginning of iterator is smaller than requested, or because
the kernel failed to pin that many pages.

direct-io.c switched to using iov_iter_get_pages()
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

fs/direct-io.c

@@ -77,7 +77,6 @@ struct dio_submit {
 	unsigned blocks_available;	/* At block_in_file.  changes */
 	int reap_counter;		/* rate limit reaping */
 	sector_t final_block_in_request;/* doesn't change */
-	unsigned first_block_in_page;	/* doesn't change, Used only once */
 	int boundary;			/* prev block is at a boundary */
 	get_block_t *get_block;		/* block mapping function */
 	dio_submit_t *submit_io;	/* IO submition function */
@@ -98,19 +97,14 @@ struct dio_submit {
 	sector_t cur_page_block;	/* Where it starts */
 	loff_t cur_page_fs_offset;	/* Offset in file */
 
-	/*
-	 * Page fetching state. These variables belong to dio_refill_pages().
-	 */
-	int curr_page;			/* changes */
-	int total_pages;		/* doesn't change */
-	unsigned long curr_user_address;/* changes */
-
+	struct iov_iter *iter;
 	/*
 	 * Page queue.  These variables belong to dio_refill_pages() and
 	 * dio_get_page().
 	 */
 	unsigned head;			/* next page to process */
 	unsigned tail;			/* last valid page + 1 */
+	size_t from, to;
 };
 
 /* dio_state communicated between submission path and end_io */
@@ -163,15 +157,10 @@ static inline unsigned dio_pages_present(struct dio_submit *sdio)
  */
 static inline int dio_refill_pages(struct dio *dio, struct dio_submit *sdio)
 {
-	int ret;
-	int nr_pages;
+	ssize_t ret;
 
-	nr_pages = min(sdio->total_pages - sdio->curr_page, DIO_PAGES);
-	ret = get_user_pages_fast(
-		sdio->curr_user_address,		/* Where from? */
-		nr_pages,			/* How many pages? */
-		dio->rw == READ,		/* Write to memory? */
-		&dio->pages[0]);		/* Put results here */
+	ret = iov_iter_get_pages(sdio->iter, dio->pages, DIO_PAGES * PAGE_SIZE,
+				&sdio->from);
 
 	if (ret < 0 && sdio->blocks_available && (dio->rw & WRITE)) {
 		struct page *page = ZERO_PAGE(0);
@@ -186,18 +175,19 @@ static inline int dio_refill_pages(struct dio *dio, struct dio_submit *sdio)
 		dio->pages[0] = page;
 		sdio->head = 0;
 		sdio->tail = 1;
-		ret = 0;
-		goto out;
+		sdio->from = 0;
+		sdio->to = PAGE_SIZE;
+		return 0;
 	}
 
 	if (ret >= 0) {
-		sdio->curr_user_address += ret * PAGE_SIZE;
-		sdio->curr_page += ret;
+		iov_iter_advance(sdio->iter, ret);
+		ret += sdio->from;
 		sdio->head = 0;
-		sdio->tail = ret;
-		ret = 0;
+		sdio->tail = (ret + PAGE_SIZE - 1) / PAGE_SIZE;
+		sdio->to = ((ret - 1) & (PAGE_SIZE - 1)) + 1;
+		return 0;
 	}
-out:
 	return ret;	
 }
 
@@ -208,8 +198,9 @@ static inline int dio_refill_pages(struct dio *dio, struct dio_submit *sdio)
  * L1 cache.
  */
 static inline struct page *dio_get_page(struct dio *dio,
-		struct dio_submit *sdio)
+		struct dio_submit *sdio, size_t *from, size_t *to)
 {
+	int n;
 	if (dio_pages_present(sdio) == 0) {
 		int ret;
 
@@ -218,7 +209,10 @@ static inline struct page *dio_get_page(struct dio *dio,
 			return ERR_PTR(ret);
 		BUG_ON(dio_pages_present(sdio) == 0);
 	}
-	return dio->pages[sdio->head++];
+	n = sdio->head++;
+	*from = n ? 0 : sdio->from;
+	*to = (n == sdio->tail - 1) ? sdio->to : PAGE_SIZE;
+	return dio->pages[n];
 }
 
 /**
@@ -422,8 +416,8 @@ static inline void dio_bio_submit(struct dio *dio, struct dio_submit *sdio)
  */
 static inline void dio_cleanup(struct dio *dio, struct dio_submit *sdio)
 {
-	while (dio_pages_present(sdio))
-		page_cache_release(dio_get_page(dio, sdio));
+	while (sdio->head < sdio->tail)
+		page_cache_release(dio->pages[sdio->head++]);
 }
 
 /*
@@ -912,23 +906,18 @@ static int do_direct_IO(struct dio *dio, struct dio_submit *sdio,
 			struct buffer_head *map_bh)
 {
 	const unsigned blkbits = sdio->blkbits;
-	const unsigned blocks_per_page = PAGE_SIZE >> blkbits;
-	struct page *page;
-	unsigned block_in_page;
 	int ret = 0;
 
-	/* The I/O can start at any block offset within the first page */
-	block_in_page = sdio->first_block_in_page;
-
 	while (sdio->block_in_file < sdio->final_block_in_request) {
-		page = dio_get_page(dio, sdio);
+		struct page *page;
+		size_t from, to;
+		page = dio_get_page(dio, sdio, &from, &to);
 		if (IS_ERR(page)) {
 			ret = PTR_ERR(page);
 			goto out;
 		}
 
-		while (block_in_page < blocks_per_page) {
-			unsigned offset_in_page = block_in_page << blkbits;
+		while (from < to) {
 			unsigned this_chunk_bytes;	/* # of bytes mapped */
 			unsigned this_chunk_blocks;	/* # of blocks */
 			unsigned u;
@@ -999,10 +988,9 @@ static int do_direct_IO(struct dio *dio, struct dio_submit *sdio,
 					page_cache_release(page);
 					goto out;
 				}
-				zero_user(page, block_in_page << blkbits,
-						1 << blkbits);
+				zero_user(page, from, 1 << blkbits);
 				sdio->block_in_file++;
-				block_in_page++;
+				from += 1 << blkbits;
 				dio->result += 1 << blkbits;
 				goto next_block;
 			}
@@ -1020,7 +1008,7 @@ static int do_direct_IO(struct dio *dio, struct dio_submit *sdio,
 			 * can add to this page
 			 */
 			this_chunk_blocks = sdio->blocks_available;
-			u = (PAGE_SIZE - offset_in_page) >> blkbits;
+			u = (to - from) >> blkbits;
 			if (this_chunk_blocks > u)
 				this_chunk_blocks = u;
 			u = sdio->final_block_in_request - sdio->block_in_file;
@@ -1032,7 +1020,7 @@ static int do_direct_IO(struct dio *dio, struct dio_submit *sdio,
 			if (this_chunk_blocks == sdio->blocks_available)
 				sdio->boundary = buffer_boundary(map_bh);
 			ret = submit_page_section(dio, sdio, page,
-						  offset_in_page,
+						  from,
 						  this_chunk_bytes,
 						  sdio->next_block_for_io,
 						  map_bh);
@@ -1043,9 +1031,9 @@ static int do_direct_IO(struct dio *dio, struct dio_submit *sdio,
 			sdio->next_block_for_io += this_chunk_blocks;
 
 			sdio->block_in_file += this_chunk_blocks;
-			block_in_page += this_chunk_blocks;
+			from += this_chunk_bytes;
+			dio->result += this_chunk_bytes;
 			sdio->blocks_available -= this_chunk_blocks;
-			dio->result += this_chunk_blocks << blkbits;
 next_block:
 			BUG_ON(sdio->block_in_file > sdio->final_block_in_request);
 			if (sdio->block_in_file == sdio->final_block_in_request)
@@ -1054,7 +1042,6 @@ static int do_direct_IO(struct dio *dio, struct dio_submit *sdio,
 
 		/* Drop the ref which was taken in get_user_pages() */
 		page_cache_release(page);
-		block_in_page = 0;
 	}
 out:
 	return ret;
@@ -1122,7 +1109,6 @@ do_blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
 	struct dio *dio;
 	struct dio_submit sdio = { 0, };
 	unsigned long user_addr;
-	size_t bytes;
 	struct buffer_head map_bh = { 0, };
 	struct blk_plug plug;
 	unsigned long align = offset | iov_iter_alignment(iter);
@@ -1234,6 +1220,10 @@ do_blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
 	spin_lock_init(&dio->bio_lock);
 	dio->refcount = 1;
 
+	sdio.iter = iter;
+	sdio.final_block_in_request =
+		(offset + iov_iter_count(iter)) >> blkbits;
+
 	/*
 	 * In case of non-aligned buffers, we may need 2 more
 	 * pages since we need to zero out first and last block.
@@ -1250,34 +1240,9 @@ do_blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
 
 	blk_start_plug(&plug);
 
-	for (seg = 0; seg < iter->nr_segs; seg++) {
-		user_addr = (unsigned long)iter->iov[seg].iov_base;
-		sdio.size += bytes = iter->iov[seg].iov_len;
-
-		/* Index into the first page of the first block */
-		sdio.first_block_in_page = (user_addr & ~PAGE_MASK) >> blkbits;
-		sdio.final_block_in_request = sdio.block_in_file +
-						(bytes >> blkbits);
-		/* Page fetching state */
-		sdio.head = 0;
-		sdio.tail = 0;
-		sdio.curr_page = 0;
-
-		sdio.total_pages = 0;
-		if (user_addr & (PAGE_SIZE-1)) {
-			sdio.total_pages++;
-			bytes -= PAGE_SIZE - (user_addr & (PAGE_SIZE - 1));
-		}
-		sdio.total_pages += (bytes + PAGE_SIZE - 1) / PAGE_SIZE;
-		sdio.curr_user_address = user_addr;
-
 	retval = do_direct_IO(dio, &sdio, &map_bh);
-
-		if (retval) {
+	if (retval)
 		dio_cleanup(dio, &sdio);
-			break;
-		}
-	} /* end iovec loop */
 
 	if (retval == -ENOTBLK) {
 		/*

include/linux/uio.h

@@ -71,6 +71,8 @@ size_t copy_page_to_iter(struct page *page, size_t offset, size_t bytes,
 unsigned long iov_iter_alignment(const struct iov_iter *i);
 void iov_iter_init(struct iov_iter *i, int direction, const struct iovec *iov,
 			unsigned long nr_segs, size_t count);
+ssize_t iov_iter_get_pages(struct iov_iter *i, struct page **pages,
+			size_t maxsize, size_t *start);
 
 static inline size_t iov_iter_count(struct iov_iter *i)
 {

mm/iov_iter.c

@@ -235,3 +235,30 @@ void iov_iter_init(struct iov_iter *i, int direction,
 	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);