cleanup for the recent bio_iov_iter_get_pages changes

cleanup for the recent bio_iov_iter_get_pages changes

[Christoph Hellwig]

Hi all,

while looking over the bio splitting issue reported by Qu, I noticed
that some of the recent changes to bio_iov_iter_get_pages lead to
more indirections than really needed, especially with the bcachefs
abuse now removed in 6.18-rc.  This small series cleans this up
an prepares for the file system block size splitting needed by
btrfs bs > PAGE_SIZE support.

Diffstat:
 block/bio.c            |    5 ++---
 block/blk-map.c        |    6 +++++-
 block/fops.c           |   13 ++++++++++---
 fs/iomap/direct-io.c   |    3 ++-
 include/linux/bio.h    |    7 +------
 include/linux/blkdev.h |    7 -------
 6 files changed, 20 insertions(+), 21 deletions(-)

[PATCH 1/4] block: remove bio_iov_iter_get_pages

[Christoph Hellwig]

Switch the only caller to bio_iov_iter_get_pages, and explain why it does
not have any alignment requirements.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-map.c     | 6 +++++-
 include/linux/bio.h | 5 -----
 2 files changed, 5 insertions(+), 6 deletions(-)

diff --git a/block/blk-map.c b/block/blk-map.c
index 165f2234f00f..6cce652c7fa6 100644
--- a/block/blk-map.c
+++ b/block/blk-map.c
@@ -283,7 +283,11 @@ static int bio_map_user_iov(struct request *rq, struct iov_iter *iter,
 	bio = blk_rq_map_bio_alloc(rq, nr_vecs, gfp_mask);
 	if (!bio)
 		return -ENOMEM;
-	ret = bio_iov_iter_get_pages(bio, iter);
+	/*
+	 * No alignment requirements on our part to support arbitrary
+	 * passthrough commands.
+	 */
+	ret = bio_iov_iter_get_pages_aligned(bio, iter, 0);
 	if (ret)
 		goto out_put;
 	ret = blk_rq_append_bio(rq, bio);
diff --git a/include/linux/bio.h b/include/linux/bio.h
index a64a30131031..b01dae9506de 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -449,11 +449,6 @@ int bdev_rw_virt(struct block_device *bdev, sector_t sector, void *data,
 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

[PATCH 2/4] block: rename bio_iov_iter_get_pages_aligned to bio_iov_iter_get_pages

[Christoph Hellwig]

Now that the bio_iov_iter_get_pages is free again, use it instead of
the more complicated now.  Also drop the unused export.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/bio.c            | 5 ++---
 block/blk-map.c        | 2 +-
 include/linux/bio.h    | 2 +-
 include/linux/blkdev.h | 2 +-
 4 files changed, 5 insertions(+), 6 deletions(-)

diff --git a/block/bio.c b/block/bio.c
index 3a1a848940dd..b3a79285c278 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -1316,7 +1316,7 @@ static int bio_iov_iter_align_down(struct bio *bio, struct iov_iter *iter,
 }
 
 /**
- * bio_iov_iter_get_pages_aligned - add user or kernel pages to a bio
+ * bio_iov_iter_get_pages - 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 the total size to, 0 for any length
@@ -1336,7 +1336,7 @@ static int bio_iov_iter_align_down(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_aligned(struct bio *bio, struct iov_iter *iter,
+int bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter,
 			   unsigned len_align_mask)
 {
 	int ret = 0;
@@ -1360,7 +1360,6 @@ int bio_iov_iter_get_pages_aligned(struct bio *bio, struct iov_iter *iter,
 		return bio_iov_iter_align_down(bio, iter, len_align_mask);
 	return ret;
 }
-EXPORT_SYMBOL_GPL(bio_iov_iter_get_pages_aligned);
 
 static void submit_bio_wait_endio(struct bio *bio)
 {
diff --git a/block/blk-map.c b/block/blk-map.c
index 6cce652c7fa6..60faf036fb6e 100644
--- a/block/blk-map.c
+++ b/block/blk-map.c
@@ -287,7 +287,7 @@ static int bio_map_user_iov(struct request *rq, struct iov_iter *iter,
 	 * No alignment requirements on our part to support arbitrary
 	 * passthrough commands.
 	 */
-	ret = bio_iov_iter_get_pages_aligned(bio, iter, 0);
+	ret = bio_iov_iter_get_pages(bio, iter, 0);
 	if (ret)
 		goto out_put;
 	ret = blk_rq_append_bio(rq, bio);
diff --git a/include/linux/bio.h b/include/linux/bio.h
index b01dae9506de..16c1c85613b7 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -446,7 +446,7 @@ 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_aligned(struct bio *bio, struct iov_iter *iter,
+int bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter,
 		unsigned len_align_mask);
 
 void bio_iov_bvec_set(struct bio *bio, const struct iov_iter *iter);
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index dad5cb5b3812..134f974ac92d 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -1876,7 +1876,7 @@ static inline int bio_split_rw_at(struct bio *bio,
 static inline int bio_iov_iter_get_bdev_pages(struct bio *bio,
 		struct iov_iter *iter, struct block_device *bdev)
 {
-	return bio_iov_iter_get_pages_aligned(bio, iter,
+	return bio_iov_iter_get_pages(bio, iter,
 					bdev_logical_block_size(bdev) - 1);
 }
 
-- 
2.47.3

[PATCH 3/4] iomap: open code bio_iov_iter_get_bdev_pages

[Christoph Hellwig]

Prepare for passing different alignments, and to retired
bio_iov_iter_get_bdev_pages as a global helper.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 fs/iomap/direct-io.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/fs/iomap/direct-io.c b/fs/iomap/direct-io.c
index 9802b2cc29bb..5d5d63efbd57 100644
--- a/fs/iomap/direct-io.c
+++ b/fs/iomap/direct-io.c
@@ -433,7 +433,8 @@ 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_bdev_pages(bio, dio->submit.iter, iomap->bdev);
+		ret = bio_iov_iter_get_pages(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
-- 
2.47.3

[PATCH 4/4] block: move bio_iov_iter_get_bdev_pages to block/fops.c

[Christoph Hellwig]

Keep bio_iov_iter_get_bdev_pages local with the callers, as blindly
looking at the bdev logical block size is often not the best idea
unless on a block device.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/fops.c           | 13 ++++++++++---
 include/linux/blkdev.h |  7 -------
 2 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/block/fops.c b/block/fops.c
index c2c0396ea9ee..5e3db9fead77 100644
--- a/block/fops.c
+++ b/block/fops.c
@@ -43,6 +43,13 @@ static bool blkdev_dio_invalid(struct block_device *bdev, struct kiocb *iocb,
 			(bdev_logical_block_size(bdev) - 1);
 }
 
+static inline int blkdev_iov_iter_get_pages(struct bio *bio,
+		struct iov_iter *iter, struct block_device *bdev)
+{
+	return bio_iov_iter_get_pages(bio, iter,
+			bdev_logical_block_size(bdev) - 1);
+}
+
 #define DIO_INLINE_BIO_VECS 4
 
 static ssize_t __blkdev_direct_IO_simple(struct kiocb *iocb,
@@ -78,7 +85,7 @@ 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_bdev_pages(&bio, iter, bdev);
+	ret = blkdev_iov_iter_get_pages(&bio, iter, bdev);
 	if (unlikely(ret))
 		goto out;
 	ret = bio.bi_iter.bi_size;
@@ -212,7 +219,7 @@ 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_bdev_pages(bio, iter, bdev);
+		ret = blkdev_iov_iter_get_pages(bio, iter, bdev);
 		if (unlikely(ret)) {
 			bio->bi_status = BLK_STS_IOERR;
 			bio_endio(bio);
@@ -348,7 +355,7 @@ static ssize_t __blkdev_direct_IO_async(struct kiocb *iocb,
 		 */
 		bio_iov_bvec_set(bio, iter);
 	} else {
-		ret = bio_iov_iter_get_bdev_pages(bio, iter, bdev);
+		ret = blkdev_iov_iter_get_pages(bio, iter, bdev);
 		if (unlikely(ret))
 			goto out_bio_put;
 	}
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 134f974ac92d..70b671a9a7f7 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -1873,13 +1873,6 @@ static inline int bio_split_rw_at(struct bio *bio,
 	return bio_split_io_at(bio, lim, segs, max_bytes, lim->dma_alignment);
 }
 
-static inline int bio_iov_iter_get_bdev_pages(struct bio *bio,
-		struct iov_iter *iter, struct block_device *bdev)
-{
-	return bio_iov_iter_get_pages(bio, iter,
-					bdev_logical_block_size(bdev) - 1);
-}
-
 #define DEFINE_IO_COMP_BATCH(name)	struct io_comp_batch name = { }
 
 #endif /* _LINUX_BLKDEV_H */
-- 
2.47.3

Re: cleanup for the recent bio_iov_iter_get_pages changes

[Qu Wenruo]

在 2025/10/7 19:36, Christoph Hellwig 写道:
> Hi all,
> 
> while looking over the bio splitting issue reported by Qu, I noticed
> that some of the recent changes to bio_iov_iter_get_pages lead to
> more indirections than really needed, especially with the bcachefs
> abuse now removed in 6.18-rc.  This small series cleans this up
> an prepares for the file system block size splitting needed by
> btrfs bs > PAGE_SIZE support.

Looks good to me, and makes my later changes much easier.

Reviewed-by: Qu Wenruo <wqu@suse.com>

Thanks,
Qu
> 
> Diffstat:
>   block/bio.c            |    5 ++---
>   block/blk-map.c        |    6 +++++-
>   block/fops.c           |   13 ++++++++++---
>   fs/iomap/direct-io.c   |    3 ++-
>   include/linux/bio.h    |    7 +------
>   include/linux/blkdev.h |    7 -------
>   6 files changed, 20 insertions(+), 21 deletions(-)

Re: [PATCH 4/4] block: move bio_iov_iter_get_bdev_pages to block/fops.c

[shinichiro.kawasaki]

[-- Attachment #1: Type: text/plain, Size: 354 bytes --]

Dear patch submitter,

Blktests CI has tested the following submission:
Status:     FAILURE
Name:       [4/4] block: move bio_iov_iter_get_bdev_pages to block/fops.c
Patchwork:  https://patchwork.kernel.org/project/linux-block/list/?series=1008963&state=*
Run record: https://github.com/linux-blktests/linux-block/actions/runs/18307874888


No failure

Re: cleanup for the recent bio_iov_iter_get_pages changes

[Keith Busch]

On Tue, Oct 07, 2025 at 11:06:24AM +0200, Christoph Hellwig wrote:
> while looking over the bio splitting issue reported by Qu, I noticed
> that some of the recent changes to bio_iov_iter_get_pages lead to
> more indirections than really needed, especially with the bcachefs
> abuse now removed in 6.18-rc.  This small series cleans this up
> an prepares for the file system block size splitting needed by
> btrfs bs > PAGE_SIZE support.

Looks good to me.

Reviewed-by: Keith Busch <kbusch@kernel.org>

Re: cleanup for the recent bio_iov_iter_get_pages changes

[Johannes Thumshirn]

Looks good,

Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>

Re: [PATCH 4/4] block: move bio_iov_iter_get_bdev_pages to block/fops.c

[Keith Busch]

On Tue, Oct 07, 2025 at 04:57:40AM -0700, shinichiro.kawasaki@wdc.com wrote:
> Blktests CI has tested the following submission:
> Status:     FAILURE

...

> No failure

o_O

Re: cleanup for the recent bio_iov_iter_get_pages changes

[Martin K. Petersen]

Christoph,

> while looking over the bio splitting issue reported by Qu, I noticed
> that some of the recent changes to bio_iov_iter_get_pages lead to more
> indirections than really needed, especially with the bcachefs abuse
> now removed in 6.18-rc. This small series cleans this up an prepares
> for the file system block size splitting needed by btrfs bs >
> PAGE_SIZE support.

Looks fine.

Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>

-- 
Martin K. Petersen

Re: cleanup for the recent bio_iov_iter_get_pages changes

[Jens Axboe]

On Tue, 07 Oct 2025 11:06:24 +0200, Christoph Hellwig wrote:
> while looking over the bio splitting issue reported by Qu, I noticed
> that some of the recent changes to bio_iov_iter_get_pages lead to
> more indirections than really needed, especially with the bcachefs
> abuse now removed in 6.18-rc.  This small series cleans this up
> an prepares for the file system block size splitting needed by
> btrfs bs > PAGE_SIZE support.
> 
> [...]

Applied, thanks!

[1/4] block: remove bio_iov_iter_get_pages
      commit: 1ed06c83506ecaaf1836ddeb7c65772ff86d8d53
[2/4] block: rename bio_iov_iter_get_pages_aligned to bio_iov_iter_get_pages
      commit: 82dd5d763c9b718e2d655b9565e0a06a91bb83dc
[3/4] iomap: open code bio_iov_iter_get_bdev_pages
      commit: cb6d51a4115781fd9de6108932e866a332e38406
[4/4] block: move bio_iov_iter_get_bdev_pages to block/fops.c
      commit: 506aa235f6e0baa00bf792df82a5e9f618b7a5d8

Best regards,
-- 
Jens Axboe

Re: [PATCH 4/4] block: move bio_iov_iter_get_bdev_pages to block/fops.c

[Shinichiro Kawasaki]

On Oct 07, 2025 / 06:57, Keith Busch wrote:
> On Tue, Oct 07, 2025 at 04:57:40AM -0700, shinichiro.kawasaki@wdc.com wrote:
> > Blktests CI has tested the following submission:
> > Status:     FAILURE
> 
> ...
> 
> > No failure
> 
> o_O

Sorry about the noise and the confusing report. I enabled the blktests CI
automated testing system, but it didn't go well. I have stopped the system,
and will work on the fix.