cleanup and refactor BLOCK_PC mapping helpers V2

cleanup and refactor BLOCK_PC mapping helpers V2

[Christoph Hellwig]

This series is based on the first two patches from the reent
series that Dongsu Park recently sent, and refactors and cleanups
up various aspect of our BLOCK_PC mapping helpers.

Changes since V1:
 - keep a separate blk_rq_map_user for now, it contains magic
   for the sg and tape drivers that's still missing from
   blk_rq_map_user_iov

[PATCH 1/7] block: simplify bio_map_kern

[Christoph Hellwig]

Just open code the trivial mapping from a kernel virtual address to
a bio instead of going through the complex user address mapping
machinery.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/bio.c | 60 ++++++++++++++++++++++++++++++++++++++++++++----------------
 1 file changed, 44 insertions(+), 16 deletions(-)

diff --git a/block/bio.c b/block/bio.c
index 471d738..54da51e 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -1563,9 +1563,8 @@ static void bio_copy_kern_endio(struct bio *bio, int err)
 {
 	struct bio_vec *bvec;
 	const int read = bio_data_dir(bio) == READ;
-	struct bio_map_data *bmd = bio->bi_private;
+	char *p = bio->bi_private;
 	int i;
-	char *p = bmd->sgvecs[0].iov_base;
 
 	bio_for_each_segment_all(bvec, bio, i) {
 		char *addr = page_address(bvec->bv_page);
@@ -1577,7 +1576,6 @@ static void bio_copy_kern_endio(struct bio *bio, int err)
 		p += bvec->bv_len;
 	}
 
-	kfree(bmd);
 	bio_put(bio);
 }
 
@@ -1595,28 +1593,58 @@ static void bio_copy_kern_endio(struct bio *bio, int err)
 struct bio *bio_copy_kern(struct request_queue *q, void *data, unsigned int len,
 			  gfp_t gfp_mask, int reading)
 {
-	struct bio *bio;
+	unsigned long kaddr = (unsigned long)data;
+	unsigned long end = (kaddr + len + PAGE_SIZE - 1) >> PAGE_SHIFT;
+	unsigned long start = kaddr >> PAGE_SHIFT;
 	struct bio_vec *bvec;
-	int i;
+	struct bio *bio;
+	void *p = data;
+	int nr_pages = 0, i;
 
-	bio = bio_copy_user(q, NULL, (unsigned long)data, len, 1, gfp_mask);
-	if (IS_ERR(bio))
-		return bio;
+	/*
+	 * Overflow, abort
+	 */
+	if (end < start)
+		return ERR_PTR(-EINVAL);
 
-	if (!reading) {
-		void *p = data;
+	nr_pages = end - start;
+	bio = bio_kmalloc(gfp_mask, nr_pages);
+	if (!bio)
+		return ERR_PTR(-ENOMEM);
 
-		bio_for_each_segment_all(bvec, bio, i) {
-			char *addr = page_address(bvec->bv_page);
+	while (len) {
+		struct page *page;
+		unsigned int bytes = PAGE_SIZE;
 
-			memcpy(addr, p, bvec->bv_len);
-			p += bvec->bv_len;
-		}
+		if (bytes > len)
+			bytes = len;
+
+		page = alloc_page(q->bounce_gfp | gfp_mask);
+		if (!page)
+			goto cleanup;
+
+		if (!reading)
+			memcpy(page_address(page), p, bytes);
+
+		if (bio_add_pc_page(q, bio, page, bytes, 0) < bytes)
+			break;
+
+		len -= bytes;
+		p += bytes;
 	}
 
-	bio->bi_end_io = bio_copy_kern_endio;
+	if (!reading)
+		bio->bi_rw |= REQ_WRITE;
 
+	bio->bi_private = data;
+	bio->bi_end_io = bio_copy_kern_endio;
 	return bio;
+
+cleanup:
+	bio_for_each_segment_all(bvec, bio, i)
+		__free_page(bvec->bv_page);
+	bio_put(bio);
+	return ERR_PTR(-ENOMEM);
 }
 EXPORT_SYMBOL(bio_copy_kern);
 
-- 
1.9.1

[PATCH 2/7] block: use blk_rq_map_user_iov to implement blk_rq_map_user

[Christoph Hellwig]

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/bio.c         |  53 +-------------------
 block/blk-map.c     | 137 +++++-----------------------------------------------
 include/linux/bio.h |   4 --
 3 files changed, 14 insertions(+), 180 deletions(-)

diff --git a/block/bio.c b/block/bio.c
index 54da51e..879921e 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -1102,7 +1102,7 @@ static int __bio_copy_iov(struct bio *bio, const struct sg_iovec *iov, int iov_c
  *	bio_uncopy_user	-	finish previously mapped bio
  *	@bio: bio being terminated
  *
- *	Free pages allocated from bio_copy_user() and write back data
+ *	Free pages allocated from bio_copy_user_iov() and write back data
  *	to user space in case of a read.
  */
 int bio_uncopy_user(struct bio *bio)
@@ -1256,32 +1256,6 @@ out_bmd:
 	return ERR_PTR(ret);
 }
 
-/**
- *	bio_copy_user	-	copy user data to bio
- *	@q: destination block queue
- *	@map_data: pointer to the rq_map_data holding pages (if necessary)
- *	@uaddr: start of user address
- *	@len: length in bytes
- *	@write_to_vm: bool indicating writing to pages or not
- *	@gfp_mask: memory allocation flags
- *
- *	Prepares and returns a bio for indirect user io, bouncing data
- *	to/from kernel pages as necessary. Must be paired with
- *	call bio_uncopy_user() on io completion.
- */
-struct bio *bio_copy_user(struct request_queue *q, struct rq_map_data *map_data,
-			  unsigned long uaddr, unsigned int len,
-			  int write_to_vm, gfp_t gfp_mask)
-{
-	struct sg_iovec iov;
-
-	iov.iov_base = (void __user *)uaddr;
-	iov.iov_len = len;
-
-	return bio_copy_user_iov(q, map_data, &iov, 1, write_to_vm, gfp_mask);
-}
-EXPORT_SYMBOL(bio_copy_user);
-
 static struct bio *__bio_map_user_iov(struct request_queue *q,
 				      struct block_device *bdev,
 				      const struct sg_iovec *iov, int iov_count,
@@ -1395,31 +1369,6 @@ static struct bio *__bio_map_user_iov(struct request_queue *q,
 }
 
 /**
- *	bio_map_user	-	map user address into bio
- *	@q: the struct request_queue for the bio
- *	@bdev: destination block device
- *	@uaddr: start of user address
- *	@len: length in bytes
- *	@write_to_vm: bool indicating writing to pages or not
- *	@gfp_mask: memory allocation flags
- *
- *	Map the user space address into a bio suitable for io to a block
- *	device. Returns an error pointer in case of error.
- */
-struct bio *bio_map_user(struct request_queue *q, struct block_device *bdev,
-			 unsigned long uaddr, unsigned int len, int write_to_vm,
-			 gfp_t gfp_mask)
-{
-	struct sg_iovec iov;
-
-	iov.iov_base = (void __user *)uaddr;
-	iov.iov_len = len;
-
-	return bio_map_user_iov(q, bdev, &iov, 1, write_to_vm, gfp_mask);
-}
-EXPORT_SYMBOL(bio_map_user);
-
-/**
  *	bio_map_user_iov - map user sg_iovec table into bio
  *	@q: the struct request_queue for the bio
  *	@bdev: destination block device
diff --git a/block/blk-map.c b/block/blk-map.c
index f890d43..152a5fe 100644
--- a/block/blk-map.c
+++ b/block/blk-map.c
@@ -39,130 +39,6 @@ static int __blk_rq_unmap_user(struct bio *bio)
 	return ret;
 }
 
-static int __blk_rq_map_user(struct request_queue *q, struct request *rq,
-			     struct rq_map_data *map_data, void __user *ubuf,
-			     unsigned int len, gfp_t gfp_mask)
-{
-	unsigned long uaddr;
-	struct bio *bio, *orig_bio;
-	int reading, ret;
-
-	reading = rq_data_dir(rq) == READ;
-
-	/*
-	 * if alignment requirement is satisfied, map in user pages for
-	 * direct dma. else, set up kernel bounce buffers
-	 */
-	uaddr = (unsigned long) ubuf;
-	if (blk_rq_aligned(q, uaddr, len) && !map_data)
-		bio = bio_map_user(q, NULL, uaddr, len, reading, gfp_mask);
-	else
-		bio = bio_copy_user(q, map_data, uaddr, len, reading, gfp_mask);
-
-	if (IS_ERR(bio))
-		return PTR_ERR(bio);
-
-	if (map_data && map_data->null_mapped)
-		bio->bi_flags |= (1 << BIO_NULL_MAPPED);
-
-	orig_bio = bio;
-	blk_queue_bounce(q, &bio);
-
-	/*
-	 * We link the bounce buffer in and could have to traverse it
-	 * later so we have to get a ref to prevent it from being freed
-	 */
-	bio_get(bio);
-
-	ret = blk_rq_append_bio(q, rq, bio);
-	if (!ret)
-		return bio->bi_iter.bi_size;
-
-	/* if it was boucned we must call the end io function */
-	bio_endio(bio, 0);
-	__blk_rq_unmap_user(orig_bio);
-	bio_put(bio);
-	return ret;
-}
-
-/**
- * blk_rq_map_user - map user data to a request, for REQ_TYPE_BLOCK_PC usage
- * @q:		request queue where request should be inserted
- * @rq:		request structure to fill
- * @map_data:   pointer to the rq_map_data holding pages (if necessary)
- * @ubuf:	the user buffer
- * @len:	length of user data
- * @gfp_mask:	memory allocation flags
- *
- * Description:
- *    Data will be mapped directly for zero copy I/O, if possible. Otherwise
- *    a kernel bounce buffer is used.
- *
- *    A matching blk_rq_unmap_user() must be issued at the end of I/O, while
- *    still in process context.
- *
- *    Note: The mapped bio may need to be bounced through blk_queue_bounce()
- *    before being submitted to the device, as pages mapped may be out of
- *    reach. It's the callers responsibility to make sure this happens. The
- *    original bio must be passed back in to blk_rq_unmap_user() for proper
- *    unmapping.
- */
-int blk_rq_map_user(struct request_queue *q, struct request *rq,
-		    struct rq_map_data *map_data, void __user *ubuf,
-		    unsigned long len, gfp_t gfp_mask)
-{
-	unsigned long bytes_read = 0;
-	struct bio *bio = NULL;
-	int ret;
-
-	if (len > (queue_max_hw_sectors(q) << 9))
-		return -EINVAL;
-	if (!len)
-		return -EINVAL;
-
-	if (!ubuf && (!map_data || !map_data->null_mapped))
-		return -EINVAL;
-
-	while (bytes_read != len) {
-		unsigned long map_len, end, start;
-
-		map_len = min_t(unsigned long, len - bytes_read, BIO_MAX_SIZE);
-		end = ((unsigned long)ubuf + map_len + PAGE_SIZE - 1)
-								>> PAGE_SHIFT;
-		start = (unsigned long)ubuf >> PAGE_SHIFT;
-
-		/*
-		 * A bad offset could cause us to require BIO_MAX_PAGES + 1
-		 * pages. If this happens we just lower the requested
-		 * mapping len by a page so that we can fit
-		 */
-		if (end - start > BIO_MAX_PAGES)
-			map_len -= PAGE_SIZE;
-
-		ret = __blk_rq_map_user(q, rq, map_data, ubuf, map_len,
-					gfp_mask);
-		if (ret < 0)
-			goto unmap_rq;
-		if (!bio)
-			bio = rq->bio;
-		bytes_read += ret;
-		ubuf += ret;
-
-		if (map_data)
-			map_data->offset += ret;
-	}
-
-	if (!bio_flagged(bio, BIO_USER_MAPPED))
-		rq->cmd_flags |= REQ_COPY_USER;
-
-	return 0;
-unmap_rq:
-	blk_rq_unmap_user(bio);
-	rq->bio = NULL;
-	return ret;
-}
-EXPORT_SYMBOL(blk_rq_map_user);
-
 /**
  * blk_rq_map_user_iov - map user data to a request, for REQ_TYPE_BLOCK_PC usage
  * @q:		request queue where request should be inserted
@@ -241,6 +117,19 @@ int blk_rq_map_user_iov(struct request_queue *q, struct request *rq,
 }
 EXPORT_SYMBOL(blk_rq_map_user_iov);
 
+int blk_rq_map_user(struct request_queue *q, struct request *rq,
+		    struct rq_map_data *map_data, void __user *ubuf,
+		    unsigned long len, gfp_t gfp_mask)
+{
+	struct sg_iovec iov;
+
+	iov.iov_base = (void __user *)ubuf;
+	iov.iov_len = len;
+
+	return blk_rq_map_user_iov(q, rq, map_data, &iov, 1, len, gfp_mask);
+}
+EXPORT_SYMBOL(blk_rq_map_user);
+
 /**
  * blk_rq_unmap_user - unmap a request with user data
  * @bio:	       start of bio list
diff --git a/include/linux/bio.h b/include/linux/bio.h
index efead0b..d0d6735 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -428,8 +428,6 @@ extern int bio_add_page(struct bio *, struct page *, unsigned int,unsigned int);
 extern int bio_add_pc_page(struct request_queue *, struct bio *, struct page *,
 			   unsigned int, unsigned int);
 extern int bio_get_nr_vecs(struct block_device *);
-extern struct bio *bio_map_user(struct request_queue *, struct block_device *,
-				unsigned long, unsigned int, int, gfp_t);
 struct sg_iovec;
 struct rq_map_data;
 extern struct bio *bio_map_user_iov(struct request_queue *,
@@ -462,8 +460,6 @@ static inline void bio_flush_dcache_pages(struct bio *bi)
 extern void bio_copy_data(struct bio *dst, struct bio *src);
 extern int bio_alloc_pages(struct bio *bio, gfp_t gfp);
 
-extern struct bio *bio_copy_user(struct request_queue *, struct rq_map_data *,
-				 unsigned long, unsigned int, int, gfp_t);
 extern struct bio *bio_copy_user_iov(struct request_queue *,
 				     struct rq_map_data *,
 				     const struct sg_iovec *,
-- 
1.9.1

[PATCH 3/7] block: add a helper to free bio bounce buffer pages

[Christoph Hellwig]

The code sniplet to walk all bio_vecs and free their pages is opencoded in
way to many places, so factor it into a helper.  Also convert the slightly
more complex cases in bio_kern_endio and __bio_copy_iov where we break
the freeing from an existing loop into a separate one.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/bio.c | 65 +++++++++++++++++++++++++++++++------------------------------
 1 file changed, 33 insertions(+), 32 deletions(-)

diff --git a/block/bio.c b/block/bio.c
index 879921e..0895f69 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -1048,7 +1048,7 @@ static struct bio_map_data *bio_alloc_map_data(unsigned int iov_count,
 }
 
 static int __bio_copy_iov(struct bio *bio, const struct sg_iovec *iov, int iov_count,
-			  int to_user, int from_user, int do_free_page)
+                          int to_user, int from_user)
 {
 	int ret = 0, i;
 	struct bio_vec *bvec;
@@ -1090,14 +1090,20 @@ static int __bio_copy_iov(struct bio *bio, const struct sg_iovec *iov, int iov_c
 				iov_off = 0;
 			}
 		}
-
-		if (do_free_page)
-			__free_page(bvec->bv_page);
 	}
 
 	return ret;
 }
 
+static void bio_free_pages(struct bio *bio)
+{
+	struct bio_vec *bvec;
+	int i;
+
+	bio_for_each_segment_all(bvec, bio, i)
+		__free_page(bvec->bv_page);
+}
+
 /**
  *	bio_uncopy_user	-	finish previously mapped bio
  *	@bio: bio being terminated
@@ -1108,8 +1114,7 @@ static int __bio_copy_iov(struct bio *bio, const struct sg_iovec *iov, int iov_c
 int bio_uncopy_user(struct bio *bio)
 {
 	struct bio_map_data *bmd = bio->bi_private;
-	struct bio_vec *bvec;
-	int ret = 0, i;
+	int ret = 0;
 
 	if (!bio_flagged(bio, BIO_NULL_MAPPED)) {
 		/*
@@ -1118,11 +1123,9 @@ int bio_uncopy_user(struct bio *bio)
 		 */
 		if (current->mm)
 			ret = __bio_copy_iov(bio, bmd->sgvecs, bmd->nr_sgvecs,
-					     bio_data_dir(bio) == READ,
-					     0, bmd->is_our_pages);
-		else if (bmd->is_our_pages)
-			bio_for_each_segment_all(bvec, bio, i)
-				__free_page(bvec->bv_page);
+					     bio_data_dir(bio) == READ, 0);
+		if (bmd->is_our_pages)
+			bio_free_pages(bio);
 	}
 	kfree(bmd);
 	bio_put(bio);
@@ -1149,7 +1152,6 @@ struct bio *bio_copy_user_iov(struct request_queue *q,
 			      int write_to_vm, gfp_t gfp_mask)
 {
 	struct bio_map_data *bmd;
-	struct bio_vec *bvec;
 	struct page *page;
 	struct bio *bio;
 	int i, ret;
@@ -1238,7 +1240,7 @@ struct bio *bio_copy_user_iov(struct request_queue *q,
 	 */
 	if ((!write_to_vm && (!map_data || !map_data->null_mapped)) ||
 	    (map_data && map_data->from_user)) {
-		ret = __bio_copy_iov(bio, iov, iov_count, 0, 1, 0);
+		ret = __bio_copy_iov(bio, iov, iov_count, 0, 1);
 		if (ret)
 			goto cleanup;
 	}
@@ -1247,9 +1249,7 @@ struct bio *bio_copy_user_iov(struct request_queue *q,
 	return bio;
 cleanup:
 	if (!map_data)
-		bio_for_each_segment_all(bvec, bio, i)
-			__free_page(bvec->bv_page);
-
+		bio_free_pages(bio);
 	bio_put(bio);
 out_bmd:
 	kfree(bmd);
@@ -1510,22 +1510,22 @@ EXPORT_SYMBOL(bio_map_kern);
 
 static void bio_copy_kern_endio(struct bio *bio, int err)
 {
-	struct bio_vec *bvec;
-	const int read = bio_data_dir(bio) == READ;
+	bio_free_pages(bio);
+	bio_put(bio);
+}
+
+static void bio_copy_kern_endio_read(struct bio *bio, int err)
+{
 	char *p = bio->bi_private;
+	struct bio_vec *bvec;
 	int i;
 
 	bio_for_each_segment_all(bvec, bio, i) {
-		char *addr = page_address(bvec->bv_page);
-
-		if (read)
-			memcpy(p, addr, bvec->bv_len);
-
-		__free_page(bvec->bv_page);
+		memcpy(p, page_address(bvec->bv_page), bvec->bv_len);
 		p += bvec->bv_len;
 	}
 
-	bio_put(bio);
+	bio_copy_kern_endio(bio, err);
 }
 
 /**
@@ -1545,10 +1545,9 @@ struct bio *bio_copy_kern(struct request_queue *q, void *data, unsigned int len,
 	unsigned long kaddr = (unsigned long)data;
 	unsigned long end = (kaddr + len + PAGE_SIZE - 1) >> PAGE_SHIFT;
 	unsigned long start = kaddr >> PAGE_SHIFT;
-	struct bio_vec *bvec;
 	struct bio *bio;
 	void *p = data;
-	int nr_pages = 0, i;
+	int nr_pages = 0;
 
 	/*
 	 * Overflow, abort
@@ -1582,16 +1581,18 @@ struct bio *bio_copy_kern(struct request_queue *q, void *data, unsigned int len,
 		p += bytes;
 	}
 
-	if (!reading)
+	if (reading) {
+		bio->bi_end_io = bio_copy_kern_endio_read;
+		bio->bi_private = data;
+	} else {
+		bio->bi_end_io = bio_copy_kern_endio;
 		bio->bi_rw |= REQ_WRITE;
+	}
 
-	bio->bi_private = data;
-	bio->bi_end_io = bio_copy_kern_endio;
 	return bio;
 
 cleanup:
-	bio_for_each_segment_all(bvec, bio, i)
-		__free_page(bvec->bv_page);
+	bio_free_pages(bio);
 	bio_put(bio);
 	return ERR_PTR(-ENOMEM);
 }
-- 
1.9.1

[PATCH 4/7] block: pass iov_iter to the BLOCK_PC mapping functions

[Christoph Hellwig]

From: Kent Overstreet <kmo@daterainc.com>

Make use of a new interface provided by iov_iter, backed by
scatter-gather list of iovec, instead of the old interface based on
sg_iovec. Also use iov_iter_advance() instead of manual iteration.

This commit should contain only literal replacements, without
functional changes.

Cc: Christoph Hellwig <hch@infradead.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Doug Gilbert <dgilbert@interlog.com>
Cc: "James E.J. Bottomley" <JBottomley@parallels.com>
Signed-off-by: Kent Overstreet <kmo@daterainc.com>
[dpark: add more description in commit message]
Signed-off-by: Dongsu Park <dongsu.park@profitbricks.com>
[hch: fixed to do a deep clone of the iov_iter, and to properly use
      the iov_iter direction]
Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/bio.c            | 146 +++++++++++++++++++++++--------------------------
 block/blk-map.c        |  38 ++++++-------
 block/scsi_ioctl.c     |  17 ++----
 drivers/scsi/sg.c      |  15 ++---
 include/linux/bio.h    |   7 +--
 include/linux/blkdev.h |   4 +-
 6 files changed, 101 insertions(+), 126 deletions(-)

diff --git a/block/bio.c b/block/bio.c
index 0895f69..7d8c655 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -28,7 +28,6 @@
 #include <linux/mempool.h>
 #include <linux/workqueue.h>
 #include <linux/cgroup.h>
-#include <scsi/sg.h>		/* for struct sg_iovec */
 
 #include <trace/events/block.h>
 
@@ -1022,21 +1021,11 @@ void bio_copy_data(struct bio *dst, struct bio *src)
 EXPORT_SYMBOL(bio_copy_data);
 
 struct bio_map_data {
-	int nr_sgvecs;
 	int is_our_pages;
-	struct sg_iovec sgvecs[];
+	struct iov_iter iter;
+	struct iovec iov[];
 };
 
-static void bio_set_map_data(struct bio_map_data *bmd, struct bio *bio,
-			     const struct sg_iovec *iov, int iov_count,
-			     int is_our_pages)
-{
-	memcpy(bmd->sgvecs, iov, sizeof(struct sg_iovec) * iov_count);
-	bmd->nr_sgvecs = iov_count;
-	bmd->is_our_pages = is_our_pages;
-	bio->bi_private = bmd;
-}
-
 static struct bio_map_data *bio_alloc_map_data(unsigned int iov_count,
 					       gfp_t gfp_mask)
 {
@@ -1044,36 +1033,33 @@ static struct bio_map_data *bio_alloc_map_data(unsigned int iov_count,
 		return NULL;
 
 	return kmalloc(sizeof(struct bio_map_data) +
-		       sizeof(struct sg_iovec) * iov_count, gfp_mask);
+		       sizeof(struct iovec) * iov_count, gfp_mask);
 }
 
-static int __bio_copy_iov(struct bio *bio, const struct sg_iovec *iov, int iov_count,
-                          int to_user, int from_user)
+static int __bio_copy_iov(struct bio *bio, const struct iov_iter *iter,
+			  int to_user, int from_user)
 {
 	int ret = 0, i;
 	struct bio_vec *bvec;
-	int iov_idx = 0;
-	unsigned int iov_off = 0;
+	struct iov_iter iov_iter = *iter;
 
 	bio_for_each_segment_all(bvec, bio, i) {
 		char *bv_addr = page_address(bvec->bv_page);
 		unsigned int bv_len = bvec->bv_len;
 
-		while (bv_len && iov_idx < iov_count) {
-			unsigned int bytes;
-			char __user *iov_addr;
-
-			bytes = min_t(unsigned int,
-				      iov[iov_idx].iov_len - iov_off, bv_len);
-			iov_addr = iov[iov_idx].iov_base + iov_off;
+		while (bv_len && iov_iter.count) {
+			struct iovec iov = iov_iter_iovec(&iov_iter);
+			unsigned int bytes = min_t(unsigned int, bv_len,
+						   iov.iov_len);
 
 			if (!ret) {
 				if (to_user)
-					ret = copy_to_user(iov_addr, bv_addr,
-							   bytes);
+					ret = copy_to_user(iov.iov_base,
+							   bv_addr, bytes);
 
 				if (from_user)
-					ret = copy_from_user(bv_addr, iov_addr,
+					ret = copy_from_user(bv_addr,
+							     iov.iov_base,
 							     bytes);
 
 				if (ret)
@@ -1082,13 +1068,7 @@ static int __bio_copy_iov(struct bio *bio, const struct sg_iovec *iov, int iov_c
 
 			bv_len -= bytes;
 			bv_addr += bytes;
-			iov_addr += bytes;
-			iov_off += bytes;
-
-			if (iov[iov_idx].iov_len == iov_off) {
-				iov_idx++;
-				iov_off = 0;
-			}
+			iov_iter_advance(&iov_iter, bytes);
 		}
 	}
 
@@ -1122,7 +1102,7 @@ int bio_uncopy_user(struct bio *bio)
 		 * don't copy into a random user address space, just free.
 		 */
 		if (current->mm)
-			ret = __bio_copy_iov(bio, bmd->sgvecs, bmd->nr_sgvecs,
+			ret = __bio_copy_iov(bio, &bmd->iter,
 					     bio_data_dir(bio) == READ, 0);
 		if (bmd->is_our_pages)
 			bio_free_pages(bio);
@@ -1135,12 +1115,10 @@ EXPORT_SYMBOL(bio_uncopy_user);
 
 /**
  *	bio_copy_user_iov	-	copy user data to bio
- *	@q: destination block queue
- *	@map_data: pointer to the rq_map_data holding pages (if necessary)
- *	@iov:	the iovec.
- *	@iov_count: number of elements in the iovec
- *	@write_to_vm: bool indicating writing to pages or not
- *	@gfp_mask: memory allocation flags
+ *	@q:		destination block queue
+ *	@map_data:	pointer to the rq_map_data holding pages (if necessary)
+ *	@iter:		iovec iterator
+ *	@gfp_mask:	memory allocation flags
  *
  *	Prepares and returns a bio for indirect user io, bouncing data
  *	to/from kernel pages as necessary. Must be paired with
@@ -1148,24 +1126,25 @@ EXPORT_SYMBOL(bio_uncopy_user);
  */
 struct bio *bio_copy_user_iov(struct request_queue *q,
 			      struct rq_map_data *map_data,
-			      const struct sg_iovec *iov, int iov_count,
-			      int write_to_vm, gfp_t gfp_mask)
+			      const struct iov_iter *iter,
+			      gfp_t gfp_mask)
 {
 	struct bio_map_data *bmd;
 	struct page *page;
 	struct bio *bio;
 	int i, ret;
 	int nr_pages = 0;
-	unsigned int len = 0;
+	unsigned int len = iter->count;
 	unsigned int offset = map_data ? map_data->offset & ~PAGE_MASK : 0;
 
-	for (i = 0; i < iov_count; i++) {
+	for (i = 0; i < iter->nr_segs; i++) {
 		unsigned long uaddr;
 		unsigned long end;
 		unsigned long start;
 
-		uaddr = (unsigned long)iov[i].iov_base;
-		end = (uaddr + iov[i].iov_len + PAGE_SIZE - 1) >> PAGE_SHIFT;
+		uaddr = (unsigned long) iter->iov[i].iov_base;
+		end = (uaddr + iter->iov[i].iov_len + PAGE_SIZE - 1)
+			>> PAGE_SHIFT;
 		start = uaddr >> PAGE_SHIFT;
 
 		/*
@@ -1175,22 +1154,31 @@ struct bio *bio_copy_user_iov(struct request_queue *q,
 			return ERR_PTR(-EINVAL);
 
 		nr_pages += end - start;
-		len += iov[i].iov_len;
 	}
 
 	if (offset)
 		nr_pages++;
 
-	bmd = bio_alloc_map_data(iov_count, gfp_mask);
+	bmd = bio_alloc_map_data(iter->nr_segs, gfp_mask);
 	if (!bmd)
 		return ERR_PTR(-ENOMEM);
 
+	/*
+	 * We need to do a deep copy of the iov_iter including the iovecs.
+	 * The caller provided iov might point to an on-stack or otherwise
+	 * shortlived one.
+	 */
+	bmd->is_our_pages = map_data ? 0 : 1;
+	memcpy(bmd->iov, iter->iov, sizeof(struct iovec) * iter->nr_segs);
+	iov_iter_init(&bmd->iter, iter->type, bmd->iov,
+			iter->nr_segs, iter->count);
+
 	ret = -ENOMEM;
 	bio = bio_kmalloc(gfp_mask, nr_pages);
 	if (!bio)
 		goto out_bmd;
 
-	if (!write_to_vm)
+	if (iter->type & WRITE)
 		bio->bi_rw |= REQ_WRITE;
 
 	ret = 0;
@@ -1238,14 +1226,14 @@ struct bio *bio_copy_user_iov(struct request_queue *q,
 	/*
 	 * success
 	 */
-	if ((!write_to_vm && (!map_data || !map_data->null_mapped)) ||
+	if (((iter->type & WRITE) && (!map_data || !map_data->null_mapped)) ||
 	    (map_data && map_data->from_user)) {
-		ret = __bio_copy_iov(bio, iov, iov_count, 0, 1);
+		ret = __bio_copy_iov(bio, iter, 0, 1);
 		if (ret)
 			goto cleanup;
 	}
 
-	bio_set_map_data(bmd, bio, iov, iov_count, map_data ? 0 : 1);
+	bio->bi_private = bmd;
 	return bio;
 cleanup:
 	if (!map_data)
@@ -1258,19 +1246,21 @@ out_bmd:
 
 static struct bio *__bio_map_user_iov(struct request_queue *q,
 				      struct block_device *bdev,
-				      const struct sg_iovec *iov, int iov_count,
-				      int write_to_vm, gfp_t gfp_mask)
+				      const struct iov_iter *iter,
+				      gfp_t gfp_mask)
 {
-	int i, j;
+	int j;
 	int nr_pages = 0;
 	struct page **pages;
 	struct bio *bio;
 	int cur_page = 0;
 	int ret, offset;
+	struct iov_iter i;
+	struct iovec iov;
 
-	for (i = 0; i < iov_count; i++) {
-		unsigned long uaddr = (unsigned long)iov[i].iov_base;
-		unsigned long len = iov[i].iov_len;
+	iov_for_each(iov, i, *iter) {
+		unsigned long uaddr = (unsigned long) iov.iov_base;
+		unsigned long len = iov.iov_len;
 		unsigned long end = (uaddr + len + PAGE_SIZE - 1) >> PAGE_SHIFT;
 		unsigned long start = uaddr >> PAGE_SHIFT;
 
@@ -1300,16 +1290,17 @@ static struct bio *__bio_map_user_iov(struct request_queue *q,
 	if (!pages)
 		goto out;
 
-	for (i = 0; i < iov_count; i++) {
-		unsigned long uaddr = (unsigned long)iov[i].iov_base;
-		unsigned long len = iov[i].iov_len;
+	iov_for_each(iov, i, *iter) {
+		unsigned long uaddr = (unsigned long) iov.iov_base;
+		unsigned long len = iov.iov_len;
 		unsigned long end = (uaddr + len + PAGE_SIZE - 1) >> PAGE_SHIFT;
 		unsigned long start = uaddr >> PAGE_SHIFT;
 		const int local_nr_pages = end - start;
 		const int page_limit = cur_page + local_nr_pages;
 
 		ret = get_user_pages_fast(uaddr, local_nr_pages,
-				write_to_vm, &pages[cur_page]);
+				(iter->type & WRITE) != WRITE,
+				&pages[cur_page]);
 		if (ret < local_nr_pages) {
 			ret = -EFAULT;
 			goto out_unmap;
@@ -1349,7 +1340,7 @@ static struct bio *__bio_map_user_iov(struct request_queue *q,
 	/*
 	 * set data direction, and check if mapped pages need bouncing
 	 */
-	if (!write_to_vm)
+	if (iter->type & WRITE)
 		bio->bi_rw |= REQ_WRITE;
 
 	bio->bi_bdev = bdev;
@@ -1357,10 +1348,10 @@ static struct bio *__bio_map_user_iov(struct request_queue *q,
 	return bio;
 
  out_unmap:
-	for (i = 0; i < nr_pages; i++) {
-		if(!pages[i])
+	for (j = 0; j < nr_pages; j++) {
+		if (!pages[j])
 			break;
-		page_cache_release(pages[i]);
+		page_cache_release(pages[j]);
 	}
  out:
 	kfree(pages);
@@ -1369,25 +1360,22 @@ static struct bio *__bio_map_user_iov(struct request_queue *q,
 }
 
 /**
- *	bio_map_user_iov - map user sg_iovec table into bio
- *	@q: the struct request_queue for the bio
- *	@bdev: destination block device
- *	@iov:	the iovec.
- *	@iov_count: number of elements in the iovec
- *	@write_to_vm: bool indicating writing to pages or not
- *	@gfp_mask: memory allocation flags
+ *	bio_map_user_iov - map user iovec into bio
+ *	@q:		the struct request_queue for the bio
+ *	@bdev:		destination block device
+ *	@iter:		iovec iterator
+ *	@gfp_mask:	memory allocation flags
  *
  *	Map the user space address into a bio suitable for io to a block
  *	device. Returns an error pointer in case of error.
  */
 struct bio *bio_map_user_iov(struct request_queue *q, struct block_device *bdev,
-			     const struct sg_iovec *iov, int iov_count,
-			     int write_to_vm, gfp_t gfp_mask)
+			     const struct iov_iter *iter,
+			     gfp_t gfp_mask)
 {
 	struct bio *bio;
 
-	bio = __bio_map_user_iov(q, bdev, iov, iov_count, write_to_vm,
-				 gfp_mask);
+	bio = __bio_map_user_iov(q, bdev, iter, gfp_mask);
 	if (IS_ERR(bio))
 		return bio;
 
diff --git a/block/blk-map.c b/block/blk-map.c
index 152a5fe..30e6bb8 100644
--- a/block/blk-map.c
+++ b/block/blk-map.c
@@ -5,7 +5,7 @@
 #include <linux/module.h>
 #include <linux/bio.h>
 #include <linux/blkdev.h>
-#include <scsi/sg.h>		/* for struct sg_iovec */
+#include <linux/uio.h>
 
 #include "blk.h"
 
@@ -44,9 +44,7 @@ static int __blk_rq_unmap_user(struct bio *bio)
  * @q:		request queue where request should be inserted
  * @rq:		request to map data to
  * @map_data:   pointer to the rq_map_data holding pages (if necessary)
- * @iov:	pointer to the iovec
- * @iov_count:	number of elements in the iovec
- * @len:	I/O byte count
+ * @iter:	iovec iterator
  * @gfp_mask:	memory allocation flags
  *
  * Description:
@@ -63,20 +61,21 @@ static int __blk_rq_unmap_user(struct bio *bio)
  *    unmapping.
  */
 int blk_rq_map_user_iov(struct request_queue *q, struct request *rq,
-			struct rq_map_data *map_data, const struct sg_iovec *iov,
-			int iov_count, unsigned int len, gfp_t gfp_mask)
+			struct rq_map_data *map_data,
+			const struct iov_iter *iter, gfp_t gfp_mask)
 {
 	struct bio *bio;
-	int i, read = rq_data_dir(rq) == READ;
 	int unaligned = 0;
+	struct iov_iter i;
+	struct iovec iov;
 
-	if (!iov || iov_count <= 0)
+	if (!iter || !iter->count)
 		return -EINVAL;
 
-	for (i = 0; i < iov_count; i++) {
-		unsigned long uaddr = (unsigned long)iov[i].iov_base;
+	iov_for_each(iov, i, *iter) {
+		unsigned long uaddr = (unsigned long) iov.iov_base;
 
-		if (!iov[i].iov_len)
+		if (!iov.iov_len)
 			return -EINVAL;
 
 		/*
@@ -86,16 +85,15 @@ int blk_rq_map_user_iov(struct request_queue *q, struct request *rq,
 			unaligned = 1;
 	}
 
-	if (unaligned || (q->dma_pad_mask & len) || map_data)
-		bio = bio_copy_user_iov(q, map_data, iov, iov_count, read,
-					gfp_mask);
+	if (unaligned || (q->dma_pad_mask & iter->count) || map_data)
+		bio = bio_copy_user_iov(q, map_data, iter, gfp_mask);
 	else
-		bio = bio_map_user_iov(q, NULL, iov, iov_count, read, gfp_mask);
+		bio = bio_map_user_iov(q, NULL, iter, gfp_mask);
 
 	if (IS_ERR(bio))
 		return PTR_ERR(bio);
 
-	if (bio->bi_iter.bi_size != len) {
+	if (bio->bi_iter.bi_size != iter->count) {
 		/*
 		 * Grab an extra reference to this bio, as bio_unmap_user()
 		 * expects to be able to drop it twice as it happens on the
@@ -121,12 +119,14 @@ int blk_rq_map_user(struct request_queue *q, struct request *rq,
 		    struct rq_map_data *map_data, void __user *ubuf,
 		    unsigned long len, gfp_t gfp_mask)
 {
-	struct sg_iovec iov;
+	struct iovec iov;
+	struct iov_iter i;
 
-	iov.iov_base = (void __user *)ubuf;
+	iov.iov_base = ubuf;
 	iov.iov_len = len;
+	iov_iter_init(&i, rq_data_dir(rq), &iov, 1, len);
 
-	return blk_rq_map_user_iov(q, rq, map_data, &iov, 1, len, gfp_mask);
+	return blk_rq_map_user_iov(q, rq, map_data, &i, gfp_mask);
 }
 EXPORT_SYMBOL(blk_rq_map_user);
 
diff --git a/block/scsi_ioctl.c b/block/scsi_ioctl.c
index 28163fa..e1f71c3 100644
--- a/block/scsi_ioctl.c
+++ b/block/scsi_ioctl.c
@@ -332,7 +332,7 @@ static int sg_io(struct request_queue *q, struct gendisk *bd_disk,
 
 	ret = 0;
 	if (hdr->iovec_count) {
-		size_t iov_data_len;
+		struct iov_iter i;
 		struct iovec *iov = NULL;
 
 		ret = rw_copy_check_uvector(-1, hdr->dxferp, hdr->iovec_count,
@@ -342,20 +342,11 @@ static int sg_io(struct request_queue *q, struct gendisk *bd_disk,
 			goto out_free_cdb;
 		}
 
-		iov_data_len = ret;
-		ret = 0;
-
 		/* SG_IO howto says that the shorter of the two wins */
-		if (hdr->dxfer_len < iov_data_len) {
-			hdr->iovec_count = iov_shorten(iov,
-						       hdr->iovec_count,
-						       hdr->dxfer_len);
-			iov_data_len = hdr->dxfer_len;
-		}
+		iov_iter_init(&i, rq_data_dir(rq), iov, hdr->iovec_count,
+			      min_t(unsigned, ret, hdr->dxfer_len));
 
-		ret = blk_rq_map_user_iov(q, rq, NULL, (struct sg_iovec *) iov,
-					  hdr->iovec_count,
-					  iov_data_len, GFP_KERNEL);
+		ret = blk_rq_map_user_iov(q, rq, NULL, &i, GFP_KERNEL);
 		kfree(iov);
 	} else if (hdr->dxfer_len)
 		ret = blk_rq_map_user(q, rq, NULL, hdr->dxferp, hdr->dxfer_len,
diff --git a/drivers/scsi/sg.c b/drivers/scsi/sg.c
index b14f64c..4052e59 100644
--- a/drivers/scsi/sg.c
+++ b/drivers/scsi/sg.c
@@ -1734,22 +1734,19 @@ sg_start_req(Sg_request *srp, unsigned char *cmd)
 	}
 
 	if (iov_count) {
-		int len, size = sizeof(struct sg_iovec) * iov_count;
+		int size = sizeof(struct iovec) * iov_count;
 		struct iovec *iov;
+		struct iov_iter i;
 
 		iov = memdup_user(hp->dxferp, size);
 		if (IS_ERR(iov))
 			return PTR_ERR(iov);
 
-		len = iov_length(iov, iov_count);
-		if (hp->dxfer_len < len) {
-			iov_count = iov_shorten(iov, iov_count, hp->dxfer_len);
-			len = hp->dxfer_len;
-		}
+		iov_iter_init(&i, rw, iov, iov_count,
+			      min_t(size_t, hp->dxfer_len,
+				    iov_length(iov, iov_count)));
 
-		res = blk_rq_map_user_iov(q, rq, md, (struct sg_iovec *)iov,
-					  iov_count,
-					  len, GFP_ATOMIC);
+		res = blk_rq_map_user_iov(q, rq, md, &i, GFP_ATOMIC);
 		kfree(iov);
 	} else
 		res = blk_rq_map_user(q, rq, md, hp->dxferp,
diff --git a/include/linux/bio.h b/include/linux/bio.h
index d0d6735..0d6105b 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -428,11 +428,10 @@ extern int bio_add_page(struct bio *, struct page *, unsigned int,unsigned int);
 extern int bio_add_pc_page(struct request_queue *, struct bio *, struct page *,
 			   unsigned int, unsigned int);
 extern int bio_get_nr_vecs(struct block_device *);
-struct sg_iovec;
 struct rq_map_data;
 extern struct bio *bio_map_user_iov(struct request_queue *,
 				    struct block_device *,
-				    const struct sg_iovec *, int, int, gfp_t);
+				    const struct iov_iter *, gfp_t);
 extern void bio_unmap_user(struct bio *);
 extern struct bio *bio_map_kern(struct request_queue *, void *, unsigned int,
 				gfp_t);
@@ -462,8 +461,8 @@ extern int bio_alloc_pages(struct bio *bio, gfp_t gfp);
 
 extern struct bio *bio_copy_user_iov(struct request_queue *,
 				     struct rq_map_data *,
-				     const struct sg_iovec *,
-				     int, int, gfp_t);
+				     const struct iov_iter *,
+				     gfp_t);
 extern int bio_uncopy_user(struct bio *);
 void zero_fill_bio(struct bio *bio);
 extern struct bio_vec *bvec_alloc(gfp_t, int, unsigned long *, mempool_t *);
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 92f4b4b..9621b73 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -850,8 +850,8 @@ extern int blk_rq_map_user(struct request_queue *, struct request *,
 extern int blk_rq_unmap_user(struct bio *);
 extern int blk_rq_map_kern(struct request_queue *, struct request *, void *, unsigned int, gfp_t);
 extern int blk_rq_map_user_iov(struct request_queue *, struct request *,
-			       struct rq_map_data *, const struct sg_iovec *,
-			       int, unsigned int, gfp_t);
+			       struct rq_map_data *, const struct iov_iter *,
+			       gfp_t);
 extern int blk_execute_rq(struct request_queue *, struct gendisk *,
 			  struct request *, int);
 extern void blk_execute_rq_nowait(struct request_queue *, struct gendisk *,
-- 
1.9.1

[PATCH 5/7] block: merge __bio_map_kern into bio_map_kern

[Christoph Hellwig]

This saves a little code, and allow to simplify the error handling.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/bio.c | 50 +++++++++++++++++---------------------------------
 1 file changed, 17 insertions(+), 33 deletions(-)

diff --git a/block/bio.c b/block/bio.c
index 7d8c655..a69a9c9 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -1429,8 +1429,18 @@ static void bio_map_kern_endio(struct bio *bio, int err)
 	bio_put(bio);
 }
 
-static struct bio *__bio_map_kern(struct request_queue *q, void *data,
-				  unsigned int len, gfp_t gfp_mask)
+/**
+ *	bio_map_kern	-	map kernel address into bio
+ *	@q: the struct request_queue for the bio
+ *	@data: pointer to buffer to map
+ *	@len: length in bytes
+ *	@gfp_mask: allocation flags for bio allocation
+ *
+ *	Map the kernel address into a bio suitable for io to a block
+ *	device. Returns an error pointer in case of error.
+ */
+struct bio *bio_map_kern(struct request_queue *q, void *data, unsigned int len,
+			 gfp_t gfp_mask)
 {
 	unsigned long kaddr = (unsigned long)data;
 	unsigned long end = (kaddr + len + PAGE_SIZE - 1) >> PAGE_SHIFT;
@@ -1454,8 +1464,11 @@ static struct bio *__bio_map_kern(struct request_queue *q, void *data,
 			bytes = len;
 
 		if (bio_add_pc_page(q, bio, virt_to_page(data), bytes,
-				    offset) < bytes)
-			break;
+				    offset) < bytes) {
+			/* we don't support partial mappings */
+			bio_put(bio);
+			return ERR_PTR(-EINVAL);
+		}
 
 		data += bytes;
 		len -= bytes;
@@ -1465,35 +1478,6 @@ static struct bio *__bio_map_kern(struct request_queue *q, void *data,
 	bio->bi_end_io = bio_map_kern_endio;
 	return bio;
 }
-
-/**
- *	bio_map_kern	-	map kernel address into bio
- *	@q: the struct request_queue for the bio
- *	@data: pointer to buffer to map
- *	@len: length in bytes
- *	@gfp_mask: allocation flags for bio allocation
- *
- *	Map the kernel address into a bio suitable for io to a block
- *	device. Returns an error pointer in case of error.
- */
-struct bio *bio_map_kern(struct request_queue *q, void *data, unsigned int len,
-			 gfp_t gfp_mask)
-{
-	struct bio *bio;
-
-	bio = __bio_map_kern(q, data, len, gfp_mask);
-	if (IS_ERR(bio))
-		return bio;
-
-	if (bio->bi_iter.bi_size == len)
-		return bio;
-
-	/*
-	 * Don't support partial mappings.
-	 */
-	bio_put(bio);
-	return ERR_PTR(-EINVAL);
-}
 EXPORT_SYMBOL(bio_map_kern);
 
 static void bio_copy_kern_endio(struct bio *bio, int err)
-- 
1.9.1

[PATCH 6/7] block: merge __bio_map_user_iov into bio_map_user_iov

[Christoph Hellwig]

And also remove the unused bdev argument.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/bio.c         | 56 +++++++++++++++++++----------------------------------
 block/blk-map.c     |  2 +-
 include/linux/bio.h |  1 -
 3 files changed, 21 insertions(+), 38 deletions(-)

diff --git a/block/bio.c b/block/bio.c
index a69a9c9..0723d4c 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -1244,10 +1244,18 @@ out_bmd:
 	return ERR_PTR(ret);
 }
 
-static struct bio *__bio_map_user_iov(struct request_queue *q,
-				      struct block_device *bdev,
-				      const struct iov_iter *iter,
-				      gfp_t gfp_mask)
+/**
+ *	bio_map_user_iov - map user iovec into bio
+ *	@q:		the struct request_queue for the bio
+ *	@iter:		iovec iterator
+ *	@gfp_mask:	memory allocation flags
+ *
+ *	Map the user space address into a bio suitable for io to a block
+ *	device. Returns an error pointer in case of error.
+ */
+struct bio *bio_map_user_iov(struct request_queue *q,
+			     const struct iov_iter *iter,
+			     gfp_t gfp_mask)
 {
 	int j;
 	int nr_pages = 0;
@@ -1343,8 +1351,15 @@ static struct bio *__bio_map_user_iov(struct request_queue *q,
 	if (iter->type & WRITE)
 		bio->bi_rw |= REQ_WRITE;
 
-	bio->bi_bdev = bdev;
 	bio->bi_flags |= (1 << BIO_USER_MAPPED);
+
+	/*
+	 * subtle -- if __bio_map_user() ended up bouncing a bio,
+	 * it would normally disappear when its bi_end_io is run.
+	 * however, we need it for the unmap, so grab an extra
+	 * reference to it
+	 */
+	bio_get(bio);
 	return bio;
 
  out_unmap:
@@ -1359,37 +1374,6 @@ static struct bio *__bio_map_user_iov(struct request_queue *q,
 	return ERR_PTR(ret);
 }
 
-/**
- *	bio_map_user_iov - map user iovec into bio
- *	@q:		the struct request_queue for the bio
- *	@bdev:		destination block device
- *	@iter:		iovec iterator
- *	@gfp_mask:	memory allocation flags
- *
- *	Map the user space address into a bio suitable for io to a block
- *	device. Returns an error pointer in case of error.
- */
-struct bio *bio_map_user_iov(struct request_queue *q, struct block_device *bdev,
-			     const struct iov_iter *iter,
-			     gfp_t gfp_mask)
-{
-	struct bio *bio;
-
-	bio = __bio_map_user_iov(q, bdev, iter, gfp_mask);
-	if (IS_ERR(bio))
-		return bio;
-
-	/*
-	 * subtle -- if __bio_map_user() ended up bouncing a bio,
-	 * it would normally disappear when its bi_end_io is run.
-	 * however, we need it for the unmap, so grab an extra
-	 * reference to it
-	 */
-	bio_get(bio);
-
-	return bio;
-}
-
 static void __bio_unmap_user(struct bio *bio)
 {
 	struct bio_vec *bvec;
diff --git a/block/blk-map.c b/block/blk-map.c
index 30e6bb8..0f22911 100644
--- a/block/blk-map.c
+++ b/block/blk-map.c
@@ -88,7 +88,7 @@ int blk_rq_map_user_iov(struct request_queue *q, struct request *rq,
 	if (unaligned || (q->dma_pad_mask & iter->count) || map_data)
 		bio = bio_copy_user_iov(q, map_data, iter, gfp_mask);
 	else
-		bio = bio_map_user_iov(q, NULL, iter, gfp_mask);
+		bio = bio_map_user_iov(q, iter, gfp_mask);
 
 	if (IS_ERR(bio))
 		return PTR_ERR(bio);
diff --git a/include/linux/bio.h b/include/linux/bio.h
index 0d6105b..da3a127 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -430,7 +430,6 @@ extern int bio_add_pc_page(struct request_queue *, struct bio *, struct page *,
 extern int bio_get_nr_vecs(struct block_device *);
 struct rq_map_data;
 extern struct bio *bio_map_user_iov(struct request_queue *,
-				    struct block_device *,
 				    const struct iov_iter *, gfp_t);
 extern void bio_unmap_user(struct bio *);
 extern struct bio *bio_map_kern(struct request_queue *, void *, unsigned int,
-- 
1.9.1

[PATCH 7/7] block: rewrite and split __bio_copy_iov()

[Christoph Hellwig]

From: Dongsu Park <dongsu.park@profitbricks.com>

Rewrite __bio_copy_iov using the copy_page_{from,to}_iter helpers, and
split it into two simpler functions.

This commit should contain only literal replacements, without
functional changes.

Cc: Kent Overstreet <kmo@daterainc.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Dongsu Park <dongsu.park@profitbricks.com>
[hch: removed the __bio_copy_iov wrapper]
Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/bio.c | 90 ++++++++++++++++++++++++++++++++++++++-----------------------
 1 file changed, 56 insertions(+), 34 deletions(-)

diff --git a/block/bio.c b/block/bio.c
index 0723d4c..f66a4ea 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -1036,43 +1036,66 @@ static struct bio_map_data *bio_alloc_map_data(unsigned int iov_count,
 		       sizeof(struct iovec) * iov_count, gfp_mask);
 }
 
-static int __bio_copy_iov(struct bio *bio, const struct iov_iter *iter,
-			  int to_user, int from_user)
+/**
+ * bio_copy_from_iter - copy all pages from iov_iter to bio
+ * @bio: The &struct bio which describes the I/O as destination
+ * @iter: iov_iter as source
+ *
+ * Copy all pages from iov_iter to bio.
+ * Returns 0 on success, or error on failure.
+ */
+static int bio_copy_from_iter(struct bio *bio, struct iov_iter iter)
 {
-	int ret = 0, i;
+	int i;
 	struct bio_vec *bvec;
-	struct iov_iter iov_iter = *iter;
 
 	bio_for_each_segment_all(bvec, bio, i) {
-		char *bv_addr = page_address(bvec->bv_page);
-		unsigned int bv_len = bvec->bv_len;
-
-		while (bv_len && iov_iter.count) {
-			struct iovec iov = iov_iter_iovec(&iov_iter);
-			unsigned int bytes = min_t(unsigned int, bv_len,
-						   iov.iov_len);
-
-			if (!ret) {
-				if (to_user)
-					ret = copy_to_user(iov.iov_base,
-							   bv_addr, bytes);
-
-				if (from_user)
-					ret = copy_from_user(bv_addr,
-							     iov.iov_base,
-							     bytes);
-
-				if (ret)
-					ret = -EFAULT;
-			}
+		ssize_t ret;
 
-			bv_len -= bytes;
-			bv_addr += bytes;
-			iov_iter_advance(&iov_iter, bytes);
-		}
+		ret = copy_page_from_iter(bvec->bv_page,
+					  bvec->bv_offset,
+					  bvec->bv_len,
+					  &iter);
+
+		if (!iov_iter_count(&iter))
+			break;
+
+		if (ret < bvec->bv_len)
+			return -EFAULT;
 	}
 
-	return ret;
+	return 0;
+}
+
+/**
+ * bio_copy_to_iter - copy all pages from bio to iov_iter
+ * @bio: The &struct bio which describes the I/O as source
+ * @iter: iov_iter as destination
+ *
+ * Copy all pages from bio to iov_iter.
+ * Returns 0 on success, or error on failure.
+ */
+static int bio_copy_to_iter(struct bio *bio, struct iov_iter iter)
+{
+	int i;
+	struct bio_vec *bvec;
+
+	bio_for_each_segment_all(bvec, bio, i) {
+		ssize_t ret;
+
+		ret = copy_page_to_iter(bvec->bv_page,
+					bvec->bv_offset,
+					bvec->bv_len,
+					&iter);
+
+		if (!iov_iter_count(&iter))
+			break;
+
+		if (ret < bvec->bv_len)
+			return -EFAULT;
+	}
+
+	return 0;
 }
 
 static void bio_free_pages(struct bio *bio)
@@ -1101,9 +1124,8 @@ int bio_uncopy_user(struct bio *bio)
 		 * if we're in a workqueue, the request is orphaned, so
 		 * don't copy into a random user address space, just free.
 		 */
-		if (current->mm)
-			ret = __bio_copy_iov(bio, &bmd->iter,
-					     bio_data_dir(bio) == READ, 0);
+		if (current->mm && bio_data_dir(bio) == READ)
+			ret = bio_copy_to_iter(bio, bmd->iter);
 		if (bmd->is_our_pages)
 			bio_free_pages(bio);
 	}
@@ -1228,7 +1250,7 @@ struct bio *bio_copy_user_iov(struct request_queue *q,
 	 */
 	if (((iter->type & WRITE) && (!map_data || !map_data->null_mapped)) ||
 	    (map_data && map_data->from_user)) {
-		ret = __bio_copy_iov(bio, iter, 0, 1);
+		ret = bio_copy_from_iter(bio, *iter);
 		if (ret)
 			goto cleanup;
 	}
-- 
1.9.1

Re: cleanup and refactor BLOCK_PC mapping helpers V2

[Ming Lei]

On 01/18/2015 11:16 PM, Christoph Hellwig wrote:
> This series is based on the first two patches from the reent
> series that Dongsu Park recently sent, and refactors and cleanups
> up various aspect of our BLOCK_PC mapping helpers.
>
> Changes since V1:
>   - keep a separate blk_rq_map_user for now, it contains magic
>     for the sg and tape drivers that's still missing from
>     blk_rq_map_user_iov

Looks all are nice cleanup:

Reviewed-by: Ming Lei <tom.leiming@gmail.com>

Also patches passed xfstests(check -g auto) against v3.19-rc4_next-20150115.

Thanks,
Ming Lei

Re: cleanup and refactor BLOCK_PC mapping helpers V2

[Christoph Hellwig]

On Sun, Jan 25, 2015 at 06:29:11PM +0800, Ming Lei wrote:
> Looks all are nice cleanup:
>
> Reviewed-by: Ming Lei <tom.leiming@gmail.com>
>
> Also patches passed xfstests(check -g auto) against v3.19-rc4_next-20150115.

Jens, do these patches look fine to you?  Any chance to get them into
the tree for the 3.20 merge window?

Re: cleanup and refactor BLOCK_PC mapping helpers V2

[Christoph Hellwig]

On Mon, Feb 02, 2015 at 02:19:09PM +0100, Christoph Hellwig wrote:
> On Sun, Jan 25, 2015 at 06:29:11PM +0800, Ming Lei wrote:
> > Looks all are nice cleanup:
> >
> > Reviewed-by: Ming Lei <tom.leiming@gmail.com>
> >
> > Also patches passed xfstests(check -g auto) against v3.19-rc4_next-20150115.
> 
> Jens, do these patches look fine to you?  Any chance to get them into
> the tree for the 3.20 merge window?

ping?

Re: cleanup and refactor BLOCK_PC mapping helpers V2

[Jens Axboe]

On 02/02/2015 06:19 AM, Christoph Hellwig wrote:
> On Sun, Jan 25, 2015 at 06:29:11PM +0800, Ming Lei wrote:
>> Looks all are nice cleanup:
>>
>> Reviewed-by: Ming Lei <tom.leiming@gmail.com>
>>
>> Also patches passed xfstests(check -g auto) against v3.19-rc4_next-20150115.
>
> Jens, do these patches look fine to you?  Any chance to get them into
> the tree for the 3.20 merge window?

Yes, I think they look fine. I'll throw them into the testing mix and 
merge them for 3.20.

-- 
Jens Axboe

Re: cleanup and refactor BLOCK_PC mapping helpers V2

[Dongsu Park]

On 05.02.2015 09:28, Jens Axboe wrote:
> On 02/02/2015 06:19 AM, Christoph Hellwig wrote:
> >Jens, do these patches look fine to you?  Any chance to get them into
> >the tree for the 3.20 merge window?
> 
> Yes, I think they look fine. I'll throw them into the testing mix and merge
> them for 3.20.

Thanks a lot, and many thanks also to Christoph.

Dongsu

> -- 
> Jens Axboe
>