[PATCH v3 0/3] remove page

public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed

* [PATCH v3 0/3] remove page_endio() v3
       [not found] <CGME20230411122922eucas1p1ed50c7c4c98104f936e3057f975c72ac@eucas1p1.samsung.com>
@ 2023-04-11 12:29 ` Pankaj Raghav
  2023-04-11 12:29   ` [PATCH v3 1/3] orangefs: use folios in orangefs_readahead Pankaj Raghav
                     ` (2 more replies)
  0 siblings, 3 replies; 5+ messages in thread
From: Pankaj Raghav @ 2023-04-11 12:29 UTC (permalink / raw)
  To: hubcap, brauner, martin, willy, hch, minchan, viro, axboe, akpm,
	senozhatsky
  Cc: linux-kernel, linux-mm, devel, linux-fsdevel, linux-block,
	gost.dev, mcgrof, Pankaj Raghav

It was decided to remove the page_endio() as per the previous RFC
discussion[1] of this series and move that functionality into the caller
itself. One of the side benefit of doing that is the callers have been
modified to directly work on folios as page_endio() already worked on
folios.

As Christoph is doing ZRAM cleanups[4] which will get rid of
page_endio() function usage, I removed the final patch that removes
page_endio()[5]. I will send it separately after rc-1 once the zram
cleanups are merged.

mpage changes were tested with a simple boot testing and running a fio
workload on ext2 filesystem. orangefs was tested by Mike Marshall
(No code changes since he tested).

Changes since v2:
- Dropped the zram patch
- Dropped the patch that removes page_endio() function from filemap
- Also split mpage_submit_bio into read and write counterparts (Christoph)

Changes since v1:
- Always chain the IO to the parent as it can never be NULL (Minchan)
- Added reviewed and tested by tags

Changes since RFC 2[2]:
- Call bio_put in zram bio end io handler (Still not Acked by hch[3])
- Call folio_set_error in mpage read endio error path (Willy)
- Directly call folio->mapping in mpage write endio error path (Willy)

[1] https://lore.kernel.org/linux-mm/ZBHcl8Pz2ULb4RGD@infradead.org/
[2] https://lore.kernel.org/linux-mm/20230322135013.197076-1-p.raghav@samsung.com/
[3] https://lore.kernel.org/linux-mm/8adb0770-6124-e11f-2551-6582db27ed32@samsung.com/
[4] https://lore.kernel.org/linux-block/20230404150536.2142108-1-hch@lst.de/T/#t
[5] https://lore.kernel.org/lkml/20230403132221.94921-6-p.raghav@samsung.com/

Pankaj Raghav (3):
  orangefs: use folios in orangefs_readahead
  mpage: split submit_bio and bio end_io handler for reads and writes
  mpage: use folios in bio end_io handler

 fs/mpage.c          | 66 +++++++++++++++++++++++++++++++--------------
 fs/orangefs/inode.c |  9 ++++---
 2 files changed, 51 insertions(+), 24 deletions(-)

-- 
2.34.1


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH v3 1/3] orangefs: use folios in orangefs_readahead
  2023-04-11 12:29 ` [PATCH v3 0/3] remove page_endio() v3 Pankaj Raghav
@ 2023-04-11 12:29   ` Pankaj Raghav
  2023-04-11 12:29   ` [PATCH v3 2/3] mpage: split submit_bio and bio end_io handler for reads and writes Pankaj Raghav
  2023-04-11 12:29   ` [PATCH v3 3/3] mpage: use folios in bio end_io handler Pankaj Raghav
  2 siblings, 0 replies; 5+ messages in thread
From: Pankaj Raghav @ 2023-04-11 12:29 UTC (permalink / raw)
  To: hubcap, brauner, martin, willy, hch, minchan, viro, axboe, akpm,
	senozhatsky
  Cc: linux-kernel, linux-mm, devel, linux-fsdevel, linux-block,
	gost.dev, mcgrof, Pankaj Raghav

Convert orangefs_readahead() from using struct page to struct folio.
This conversion removes the call to page_endio() which is soon to be
removed, and simplifies the final page handling.

The page error flags is not required to be set in the error case as
orangefs doesn't depend on them.

Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Tested-by: Mike Marshall <hubcap@omnibond.com>
Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
---
 fs/orangefs/inode.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/fs/orangefs/inode.c b/fs/orangefs/inode.c
index aefdf1d3be7c..9014bbcc8031 100644
--- a/fs/orangefs/inode.c
+++ b/fs/orangefs/inode.c
@@ -244,7 +244,7 @@ static void orangefs_readahead(struct readahead_control *rac)
 	struct iov_iter iter;
 	struct inode *inode = rac->mapping->host;
 	struct xarray *i_pages;
-	struct page *page;
+	struct folio *folio;
 	loff_t new_start = readahead_pos(rac);
 	int ret;
 	size_t new_len = 0;
@@ -275,9 +275,10 @@ static void orangefs_readahead(struct readahead_control *rac)
 		ret = 0;
 
 	/* clean up. */
-	while ((page = readahead_page(rac))) {
-		page_endio(page, false, ret);
-		put_page(page);
+	while ((folio = readahead_folio(rac))) {
+		if (!ret)
+			folio_mark_uptodate(folio);
+		folio_unlock(folio);
 	}
 }
 
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 5+ messages in thread

* [PATCH v3 2/3] mpage: split submit_bio and bio end_io handler for reads and writes
  2023-04-11 12:29 ` [PATCH v3 0/3] remove page_endio() v3 Pankaj Raghav
  2023-04-11 12:29   ` [PATCH v3 1/3] orangefs: use folios in orangefs_readahead Pankaj Raghav
@ 2023-04-11 12:29   ` Pankaj Raghav
  2023-04-11 12:30     ` Christoph Hellwig
  2023-04-11 12:29   ` [PATCH v3 3/3] mpage: use folios in bio end_io handler Pankaj Raghav
  2 siblings, 1 reply; 5+ messages in thread
From: Pankaj Raghav @ 2023-04-11 12:29 UTC (permalink / raw)
  To: hubcap, brauner, martin, willy, hch, minchan, viro, axboe, akpm,
	senozhatsky
  Cc: linux-kernel, linux-mm, devel, linux-fsdevel, linux-block,
	gost.dev, mcgrof, Pankaj Raghav

Split the submit_bio() and bio end_io handler for reads and writes similar
to other aops.
This is a prep patch before we convert end_io handlers to use folios.

Suggested-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
---
 fs/mpage.c | 54 ++++++++++++++++++++++++++++++++++++------------------
 1 file changed, 36 insertions(+), 18 deletions(-)

diff --git a/fs/mpage.c b/fs/mpage.c
index 22b9de5ddd68..d9540c1b7427 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -43,23 +43,41 @@
  * status of that page is hard.  See end_buffer_async_read() for the details.
  * There is no point in duplicating all that complexity.
  */
-static void mpage_end_io(struct bio *bio)
+static void mpage_read_end_io(struct bio *bio)
 {
 	struct bio_vec *bv;
 	struct bvec_iter_all iter_all;
 
-	bio_for_each_segment_all(bv, bio, iter_all) {
-		struct page *page = bv->bv_page;
-		page_endio(page, bio_op(bio),
+	bio_for_each_segment_all(bv, bio, iter_all)
+		page_endio(bv->bv_page, REQ_OP_READ,
 			   blk_status_to_errno(bio->bi_status));
-	}
 
 	bio_put(bio);
 }
 
-static struct bio *mpage_bio_submit(struct bio *bio)
+static void mpage_write_end_io(struct bio *bio)
 {
-	bio->bi_end_io = mpage_end_io;
+	struct bio_vec *bv;
+	struct bvec_iter_all iter_all;
+
+	bio_for_each_segment_all(bv, bio, iter_all)
+		page_endio(bv->bv_page, REQ_OP_WRITE,
+			   blk_status_to_errno(bio->bi_status));
+
+	bio_put(bio);
+}
+
+static struct bio *mpage_bio_submit_read(struct bio *bio)
+{
+	bio->bi_end_io = mpage_read_end_io;
+	guard_bio_eod(bio);
+	submit_bio(bio);
+	return NULL;
+}
+
+static struct bio *mpage_bio_submit_write(struct bio *bio)
+{
+	bio->bi_end_io = mpage_write_end_io;
 	guard_bio_eod(bio);
 	submit_bio(bio);
 	return NULL;
@@ -265,7 +283,7 @@ static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
 	 * This folio will go to BIO.  Do we need to send this BIO off first?
 	 */
 	if (args->bio && (args->last_block_in_bio != blocks[0] - 1))
-		args->bio = mpage_bio_submit(args->bio);
+		args->bio = mpage_bio_submit_read(args->bio);
 
 alloc_new:
 	if (args->bio == NULL) {
@@ -278,7 +296,7 @@ static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
 
 	length = first_hole << blkbits;
 	if (!bio_add_folio(args->bio, folio, length, 0)) {
-		args->bio = mpage_bio_submit(args->bio);
+		args->bio = mpage_bio_submit_read(args->bio);
 		goto alloc_new;
 	}
 
@@ -286,7 +304,7 @@ static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
 	nblocks = map_bh->b_size >> blkbits;
 	if ((buffer_boundary(map_bh) && relative_block == nblocks) ||
 	    (first_hole != blocks_per_page))
-		args->bio = mpage_bio_submit(args->bio);
+		args->bio = mpage_bio_submit_read(args->bio);
 	else
 		args->last_block_in_bio = blocks[blocks_per_page - 1];
 out:
@@ -294,7 +312,7 @@ static struct bio *do_mpage_readpage(struct mpage_readpage_args *args)
 
 confused:
 	if (args->bio)
-		args->bio = mpage_bio_submit(args->bio);
+		args->bio = mpage_bio_submit_read(args->bio);
 	if (!folio_test_uptodate(folio))
 		block_read_full_folio(folio, args->get_block);
 	else
@@ -356,7 +374,7 @@ void mpage_readahead(struct readahead_control *rac, get_block_t get_block)
 		args.bio = do_mpage_readpage(&args);
 	}
 	if (args.bio)
-		mpage_bio_submit(args.bio);
+		mpage_bio_submit_read(args.bio);
 }
 EXPORT_SYMBOL(mpage_readahead);
 
@@ -373,7 +391,7 @@ int mpage_read_folio(struct folio *folio, get_block_t get_block)
 
 	args.bio = do_mpage_readpage(&args);
 	if (args.bio)
-		mpage_bio_submit(args.bio);
+		mpage_bio_submit_read(args.bio);
 	return 0;
 }
 EXPORT_SYMBOL(mpage_read_folio);
@@ -577,7 +595,7 @@ static int __mpage_writepage(struct folio *folio, struct writeback_control *wbc,
 	 * This page will go to BIO.  Do we need to send this BIO off first?
 	 */
 	if (bio && mpd->last_block_in_bio != blocks[0] - 1)
-		bio = mpage_bio_submit(bio);
+		bio = mpage_bio_submit_write(bio);
 
 alloc_new:
 	if (bio == NULL) {
@@ -596,7 +614,7 @@ static int __mpage_writepage(struct folio *folio, struct writeback_control *wbc,
 	wbc_account_cgroup_owner(wbc, &folio->page, folio_size(folio));
 	length = first_unmapped << blkbits;
 	if (!bio_add_folio(bio, folio, length, 0)) {
-		bio = mpage_bio_submit(bio);
+		bio = mpage_bio_submit_write(bio);
 		goto alloc_new;
 	}
 
@@ -606,7 +624,7 @@ static int __mpage_writepage(struct folio *folio, struct writeback_control *wbc,
 	folio_start_writeback(folio);
 	folio_unlock(folio);
 	if (boundary || (first_unmapped != blocks_per_page)) {
-		bio = mpage_bio_submit(bio);
+		bio = mpage_bio_submit_write(bio);
 		if (boundary_block) {
 			write_boundary_block(boundary_bdev,
 					boundary_block, 1 << blkbits);
@@ -618,7 +636,7 @@ static int __mpage_writepage(struct folio *folio, struct writeback_control *wbc,
 
 confused:
 	if (bio)
-		bio = mpage_bio_submit(bio);
+		bio = mpage_bio_submit_write(bio);
 
 	/*
 	 * The caller has a ref on the inode, so *mapping is stable
@@ -652,7 +670,7 @@ mpage_writepages(struct address_space *mapping,
 	blk_start_plug(&plug);
 	ret = write_cache_pages(mapping, wbc, __mpage_writepage, &mpd);
 	if (mpd.bio)
-		mpage_bio_submit(mpd.bio);
+		mpage_bio_submit_write(mpd.bio);
 	blk_finish_plug(&plug);
 	return ret;
 }
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 5+ messages in thread

* Re: [PATCH v3 2/3] mpage: split submit_bio and bio end_io handler for reads and writes
  2023-04-11 12:29   ` [PATCH v3 2/3] mpage: split submit_bio and bio end_io handler for reads and writes Pankaj Raghav
@ 2023-04-11 12:30     ` Christoph Hellwig
  0 siblings, 0 replies; 5+ messages in thread
From: Christoph Hellwig @ 2023-04-11 12:30 UTC (permalink / raw)
  To: Pankaj Raghav
  Cc: hubcap, brauner, martin, willy, hch, minchan, viro, axboe, akpm,
	senozhatsky, linux-kernel, linux-mm, devel, linux-fsdevel,
	linux-block, gost.dev, mcgrof

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH v3 3/3] mpage: use folios in bio end_io handler
  2023-04-11 12:29 ` [PATCH v3 0/3] remove page_endio() v3 Pankaj Raghav
  2023-04-11 12:29   ` [PATCH v3 1/3] orangefs: use folios in orangefs_readahead Pankaj Raghav
  2023-04-11 12:29   ` [PATCH v3 2/3] mpage: split submit_bio and bio end_io handler for reads and writes Pankaj Raghav
@ 2023-04-11 12:29   ` Pankaj Raghav
  2 siblings, 0 replies; 5+ messages in thread
From: Pankaj Raghav @ 2023-04-11 12:29 UTC (permalink / raw)
  To: hubcap, brauner, martin, willy, hch, minchan, viro, axboe, akpm,
	senozhatsky
  Cc: linux-kernel, linux-mm, devel, linux-fsdevel, linux-block,
	gost.dev, mcgrof, Pankaj Raghav

Use folios in the bio end_io handler. This conversion does the appropriate
handling on the folios in the respective end_io callback and removes the
call to page_endio(), which is soon to be removed.

Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/mpage.c | 28 ++++++++++++++++++----------
 1 file changed, 18 insertions(+), 10 deletions(-)

diff --git a/fs/mpage.c b/fs/mpage.c
index d9540c1b7427..242e213ee064 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -45,24 +45,32 @@
  */
 static void mpage_read_end_io(struct bio *bio)
 {
-	struct bio_vec *bv;
-	struct bvec_iter_all iter_all;
+	struct folio_iter fi;
+	int err = blk_status_to_errno(bio->bi_status);
 
-	bio_for_each_segment_all(bv, bio, iter_all)
-		page_endio(bv->bv_page, REQ_OP_READ,
-			   blk_status_to_errno(bio->bi_status));
+	bio_for_each_folio_all(fi, bio) {
+		if (err)
+			folio_set_error(fi.folio);
+		else
+			folio_mark_uptodate(fi.folio);
+		folio_unlock(fi.folio);
+	}
 
 	bio_put(bio);
 }
 
 static void mpage_write_end_io(struct bio *bio)
 {
-	struct bio_vec *bv;
-	struct bvec_iter_all iter_all;
+	struct folio_iter fi;
+	int err = blk_status_to_errno(bio->bi_status);
 
-	bio_for_each_segment_all(bv, bio, iter_all)
-		page_endio(bv->bv_page, REQ_OP_WRITE,
-			   blk_status_to_errno(bio->bi_status));
+	bio_for_each_folio_all(fi, bio) {
+		if (err) {
+			folio_set_error(fi.folio);
+			mapping_set_error(fi.folio->mapping, err);
+		}
+		folio_end_writeback(fi.folio);
+	}
 
 	bio_put(bio);
 }
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2023-04-11 12:30 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
     [not found] <CGME20230411122922eucas1p1ed50c7c4c98104f936e3057f975c72ac@eucas1p1.samsung.com>
2023-04-11 12:29 ` [PATCH v3 0/3] remove page_endio() v3 Pankaj Raghav
2023-04-11 12:29   ` [PATCH v3 1/3] orangefs: use folios in orangefs_readahead Pankaj Raghav
2023-04-11 12:29   ` [PATCH v3 2/3] mpage: split submit_bio and bio end_io handler for reads and writes Pankaj Raghav
2023-04-11 12:30     ` Christoph Hellwig
2023-04-11 12:29   ` [PATCH v3 3/3] mpage: use folios in bio end_io handler Pankaj Raghav

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox