From: Keith Busch <kbusch@xxxxxxxxxx>
The block layer tries to align bio vectors to the block device's logical
block size. Some cases don't have a block device, or we may want to align
to something larger, which we can't derive it from the queue limits.
Have the caller specify what they want, or allow any length alignment if
nothing was specified.
Signed-off-by: Keith Busch <kbusch@xxxxxxxxxx>
---
block/bio.c | 19 +++++++++++--------
block/fops.c | 9 ++++++---
fs/iomap/direct-io.c | 3 ++-
include/linux/bio.h | 9 ++++++++-
4 files changed, 27 insertions(+), 13 deletions(-)
diff --git a/block/bio.c b/block/bio.c
index 3b371a5da159e..44286db14355f 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -1204,7 +1204,8 @@ static unsigned int get_contig_folio_len(unsigned int *num_pages,
* For a multi-segment *iter, this function only adds pages from the next
* non-empty segment of the iov iterator.
*/
-static int __bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
+static int __bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter,
+ unsigned len_align_mask)
{
iov_iter_extraction_t extraction_flags = 0;
unsigned short nr_pages = bio->bi_max_vecs - bio->bi_vcnt;
@@ -1213,7 +1214,7 @@ static int __bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
struct page **pages = (struct page **)bv;
ssize_t size;
unsigned int num_pages, i = 0;
- size_t offset, folio_offset, left, len;
+ size_t offset, folio_offset, left, len, trim;
int ret = 0;
/*
@@ -1242,8 +1243,8 @@ static int __bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
nr_pages = DIV_ROUND_UP(offset + size, PAGE_SIZE);
- if (bio->bi_bdev) {
- size_t trim = size & (bdev_logical_block_size(bio->bi_bdev) - 1);
+ trim = size & len_align_mask;
+ if (trim) {
iov_iter_revert(iter, trim);
size -= trim;
}
@@ -1298,9 +1299,10 @@ static int __bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
}
/**
- * bio_iov_iter_get_pages - add user or kernel pages to a bio
+ * bio_iov_iter_get_pages_aligned - add user or kernel pages to a bio
* @bio: bio to add pages to
* @iter: iov iterator describing the region to be added
+ * @len_align_mask: the mask to align each vector size to, 0 for any length
*
* This takes either an iterator pointing to user memory, or one pointing to
* kernel pages (BVEC iterator). If we're adding user pages, we pin them and
@@ -1317,7 +1319,8 @@ static int __bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
* MM encounters an error pinning the requested pages, it stops. Error
* is returned only if 0 pages could be pinned.
*/
-int bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
+int bio_iov_iter_get_pages_aligned(struct bio *bio, struct iov_iter *iter,
+ unsigned len_align_mask)
{
int ret = 0;
@@ -1333,12 +1336,12 @@ int bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
if (iov_iter_extract_will_pin(iter))
bio_set_flag(bio, BIO_PAGE_PINNED);
do {
- ret = __bio_iov_iter_get_pages(bio, iter);
+ ret = __bio_iov_iter_get_pages(bio, iter, len_align_mask);
} while (!ret && iov_iter_count(iter) && !bio_full(bio, 0));
return bio->bi_vcnt ? 0 : ret;
}
-EXPORT_SYMBOL_GPL(bio_iov_iter_get_pages);
+EXPORT_SYMBOL_GPL(bio_iov_iter_get_pages_aligned);
static void submit_bio_wait_endio(struct bio *bio)
{
diff --git a/block/fops.c b/block/fops.c
index 82451ac8ff25d..6d5c1e680d4a7 100644
--- a/block/fops.c
+++ b/block/fops.c
@@ -78,7 +78,8 @@ static ssize_t __blkdev_direct_IO_simple(struct kiocb *iocb,
if (iocb->ki_flags & IOCB_ATOMIC)
bio.bi_opf |= REQ_ATOMIC;
- ret = bio_iov_iter_get_pages(&bio, iter);
+ ret = bio_iov_iter_get_pages_aligned(&bio, iter,
+ bdev_logical_block_size(bdev) - 1);
if (unlikely(ret))
goto out;
ret = bio.bi_iter.bi_size;
@@ -212,7 +213,8 @@ static ssize_t __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter,
bio->bi_end_io = blkdev_bio_end_io;
bio->bi_ioprio = iocb->ki_ioprio;
- ret = bio_iov_iter_get_pages(bio, iter);
+ ret = bio_iov_iter_get_pages_aligned(bio, iter,
+ bdev_logical_block_size(bdev) - 1);
if (unlikely(ret)) {
bio->bi_status = BLK_STS_IOERR;
bio_endio(bio);
@@ -348,7 +350,8 @@ static ssize_t __blkdev_direct_IO_async(struct kiocb *iocb,
*/
bio_iov_bvec_set(bio, iter);
} else {
- ret = bio_iov_iter_get_pages(bio, iter);
+ ret = bio_iov_iter_get_pages_aligned(bio, iter,
+ bdev_logical_block_size(bdev) - 1);
if (unlikely(ret))
goto out_bio_put;
}
diff --git a/fs/iomap/direct-io.c b/fs/iomap/direct-io.c
index 6f25d4cfea9f7..213764bdee8f2 100644
--- a/fs/iomap/direct-io.c
+++ b/fs/iomap/direct-io.c
@@ -434,7 +434,8 @@ static int iomap_dio_bio_iter(struct iomap_iter *iter, struct iomap_dio *dio)
bio->bi_private = dio;
bio->bi_end_io = iomap_dio_bio_end_io;
- ret = bio_iov_iter_get_pages(bio, dio->submit.iter);
+ ret = bio_iov_iter_get_pages_aligned(bio, dio->submit.iter,
+ bdev_logical_block_size(iomap->bdev) - 1);
if (unlikely(ret)) {
/*
* We have to stop part way through an IO. We must fall
diff --git a/include/linux/bio.h b/include/linux/bio.h
index 519a1d59805f8..788a50ff319e3 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -441,7 +441,14 @@ int submit_bio_wait(struct bio *bio);
int bdev_rw_virt(struct block_device *bdev, sector_t sector, void *data,
size_t len, enum req_op op);
-int bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter);
+int bio_iov_iter_get_pages_aligned(struct bio *bio, struct iov_iter *iter,
+ unsigned len_align_mask);
+
+static inline int bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
+{
+ return bio_iov_iter_get_pages_aligned(bio, iter, 0);
+}
+
void bio_iov_bvec_set(struct bio *bio, const struct iov_iter *iter);
void __bio_release_pages(struct bio *bio, bool mark_dirty);
extern void bio_set_pages_dirty(struct bio *bio);
--
2.47.3
