stop passing a writeback_control to swap/shmem writeout v3

stop passing a writeback_control to swap/shmem writeout v3

[Christoph Hellwig]

Hi all,

this series was intended to remove the last remaining users of
AOP_WRITEPAGE_ACTIVATE after my other pending patches removed the
rest, but spectacularly failed at that.

But instead it nicely improves the code, and removes two pointers
from struct writeback_control.

Changes since v2:
 - fix up kerneldoc to not leave around the documentation for a removed
   parameter
 - add back the for_reclaim removal

Changes since v1:
 - drop to patch to be merged through a different tree or later
 - fix !CONFIG_SWAP compilation

Diffstat:
 drivers/gpu/drm/i915/gem/i915_gem_shmem.c |    3 -
 drivers/gpu/drm/ttm/ttm_backup.c          |    9 ---
 fs/fuse/file.c                            |   11 ---
 fs/nfs/write.c                            |    2 
 include/linux/shmem_fs.h                  |    5 +
 include/linux/writeback.h                 |   11 ---
 include/trace/events/btrfs.h              |    7 --
 include/trace/events/writeback.h          |    8 --
 mm/page_io.c                              |   71 +++++++++++--------------
 mm/shmem.c                                |   18 ++----
 mm/swap.h                                 |    9 ++-
 mm/vmscan.c                               |   84 +++++++++++++-----------------
 mm/zswap.c                                |    5 -
 13 files changed, 96 insertions(+), 147 deletions(-)

[PATCH 1/6] mm: split out a writeout helper from pageout

[Christoph Hellwig]

Move the code to write back swap / shmem folios into a self-contained
helper to keep prepare for refactoring it.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 mm/vmscan.c | 94 +++++++++++++++++++++++++++--------------------------
 1 file changed, 48 insertions(+), 46 deletions(-)

diff --git a/mm/vmscan.c b/mm/vmscan.c
index f8dfd2864bbf..3af99b0978b5 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -652,14 +652,55 @@ typedef enum {
 	PAGE_CLEAN,
 } pageout_t;
 
+static pageout_t writeout(struct folio *folio, struct address_space *mapping,
+		struct swap_iocb **plug, struct list_head *folio_list)
+{
+	struct writeback_control wbc = {
+		.sync_mode = WB_SYNC_NONE,
+		.nr_to_write = SWAP_CLUSTER_MAX,
+		.range_start = 0,
+		.range_end = LLONG_MAX,
+		.for_reclaim = 1,
+		.swap_plug = plug,
+	};
+	int res;
+
+	folio_set_reclaim(folio);
+
+	/*
+	 * The large shmem folio can be split if CONFIG_THP_SWAP is not enabled
+	 * or we failed to allocate contiguous swap entries.
+	 */
+	if (shmem_mapping(mapping)) {
+		if (folio_test_large(folio))
+			wbc.list = folio_list;
+		res = shmem_writeout(folio, &wbc);
+	} else {
+		res = swap_writeout(folio, &wbc);
+	}
+
+	if (res < 0)
+		handle_write_error(mapping, folio, res);
+	if (res == AOP_WRITEPAGE_ACTIVATE) {
+		folio_clear_reclaim(folio);
+		return PAGE_ACTIVATE;
+	}
+
+	/* synchronous write? */
+	if (!folio_test_writeback(folio))
+		folio_clear_reclaim(folio);
+
+	trace_mm_vmscan_write_folio(folio);
+	node_stat_add_folio(folio, NR_VMSCAN_WRITE);
+	return PAGE_SUCCESS;
+}
+
 /*
  * pageout is called by shrink_folio_list() for each dirty folio.
  */
 static pageout_t pageout(struct folio *folio, struct address_space *mapping,
 			 struct swap_iocb **plug, struct list_head *folio_list)
 {
-	int (*writeout)(struct folio *, struct writeback_control *);
-
 	/*
 	 * We no longer attempt to writeback filesystem folios here, other
 	 * than tmpfs/shmem.  That's taken care of in page-writeback.
@@ -690,51 +731,12 @@ static pageout_t pageout(struct folio *folio, struct address_space *mapping,
 		}
 		return PAGE_KEEP;
 	}
-	if (shmem_mapping(mapping))
-		writeout = shmem_writeout;
-	else if (folio_test_anon(folio))
-		writeout = swap_writeout;
-	else
-		return PAGE_ACTIVATE;
-
-	if (folio_clear_dirty_for_io(folio)) {
-		int res;
-		struct writeback_control wbc = {
-			.sync_mode = WB_SYNC_NONE,
-			.nr_to_write = SWAP_CLUSTER_MAX,
-			.range_start = 0,
-			.range_end = LLONG_MAX,
-			.for_reclaim = 1,
-			.swap_plug = plug,
-		};
 
-		/*
-		 * The large shmem folio can be split if CONFIG_THP_SWAP is
-		 * not enabled or contiguous swap entries are failed to
-		 * allocate.
-		 */
-		if (shmem_mapping(mapping) && folio_test_large(folio))
-			wbc.list = folio_list;
-
-		folio_set_reclaim(folio);
-		res = writeout(folio, &wbc);
-		if (res < 0)
-			handle_write_error(mapping, folio, res);
-		if (res == AOP_WRITEPAGE_ACTIVATE) {
-			folio_clear_reclaim(folio);
-			return PAGE_ACTIVATE;
-		}
-
-		if (!folio_test_writeback(folio)) {
-			/* synchronous write? */
-			folio_clear_reclaim(folio);
-		}
-		trace_mm_vmscan_write_folio(folio);
-		node_stat_add_folio(folio, NR_VMSCAN_WRITE);
-		return PAGE_SUCCESS;
-	}
-
-	return PAGE_CLEAN;
+	if (!shmem_mapping(mapping) && !folio_test_anon(folio))
+		return PAGE_ACTIVATE;
+	if (!folio_clear_dirty_for_io(folio))
+		return PAGE_CLEAN;
+	return writeout(folio, mapping, plug, folio_list);
 }
 
 /*
-- 
2.47.2

Re: [PATCH 1/6] mm: split out a writeout helper from pageout

[Baolin Wang]

On 2025/6/10 13:49, Christoph Hellwig wrote:
> Move the code to write back swap / shmem folios into a self-contained
> helper to keep prepare for refactoring it.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

LGTM.
Reviewed-by: Baolin Wang <baolin.wang@linux.aibaba.com>

[PATCH 2/6] mm: stop passing a writeback_control structure to shmem_writeout

[Christoph Hellwig]

shmem_writeout only needs the swap_iocb cookie and the split folio
list.  Pass those explicitly and remove the now unused list member
from struct writeback_control.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/gpu/drm/i915/gem/i915_gem_shmem.c |  2 +-
 drivers/gpu/drm/ttm/ttm_backup.c          |  9 +-------
 include/linux/shmem_fs.h                  |  5 ++++-
 include/linux/writeback.h                 |  3 ---
 mm/shmem.c                                | 26 +++++++++++++----------
 mm/vmscan.c                               | 12 +++++------
 6 files changed, 26 insertions(+), 31 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
index 19a3eb82dc6a..24d8daa4fdb3 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
@@ -317,7 +317,7 @@ void __shmem_writeback(size_t size, struct address_space *mapping)
 		if (folio_mapped(folio))
 			folio_redirty_for_writepage(&wbc, folio);
 		else
-			error = shmem_writeout(folio, &wbc);
+			error = shmem_writeout(folio, NULL, NULL);
 	}
 }
 
diff --git a/drivers/gpu/drm/ttm/ttm_backup.c b/drivers/gpu/drm/ttm/ttm_backup.c
index ffaab68bd5dd..6f2e58be4f3e 100644
--- a/drivers/gpu/drm/ttm/ttm_backup.c
+++ b/drivers/gpu/drm/ttm/ttm_backup.c
@@ -112,15 +112,8 @@ ttm_backup_backup_page(struct file *backup, struct page *page,
 
 	if (writeback && !folio_mapped(to_folio) &&
 	    folio_clear_dirty_for_io(to_folio)) {
-		struct writeback_control wbc = {
-			.sync_mode = WB_SYNC_NONE,
-			.nr_to_write = SWAP_CLUSTER_MAX,
-			.range_start = 0,
-			.range_end = LLONG_MAX,
-			.for_reclaim = 1,
-		};
 		folio_set_reclaim(to_folio);
-		ret = shmem_writeout(to_folio, &wbc);
+		ret = shmem_writeout(to_folio, NULL, NULL);
 		if (!folio_test_writeback(to_folio))
 			folio_clear_reclaim(to_folio);
 		/*
diff --git a/include/linux/shmem_fs.h b/include/linux/shmem_fs.h
index 5f03a39a26f7..6d0f9c599ff7 100644
--- a/include/linux/shmem_fs.h
+++ b/include/linux/shmem_fs.h
@@ -11,6 +11,8 @@
 #include <linux/fs_parser.h>
 #include <linux/userfaultfd_k.h>
 
+struct swap_iocb;
+
 /* inode in-kernel data */
 
 #ifdef CONFIG_TMPFS_QUOTA
@@ -107,7 +109,8 @@ static inline bool shmem_mapping(struct address_space *mapping)
 void shmem_unlock_mapping(struct address_space *mapping);
 struct page *shmem_read_mapping_page_gfp(struct address_space *mapping,
 					pgoff_t index, gfp_t gfp_mask);
-int shmem_writeout(struct folio *folio, struct writeback_control *wbc);
+int shmem_writeout(struct folio *folio, struct swap_iocb **plug,
+		struct list_head *folio_list);
 void shmem_truncate_range(struct inode *inode, loff_t start, loff_t end);
 int shmem_unuse(unsigned int type);
 
diff --git a/include/linux/writeback.h b/include/linux/writeback.h
index eda4b62511f7..82f217970092 100644
--- a/include/linux/writeback.h
+++ b/include/linux/writeback.h
@@ -79,9 +79,6 @@ struct writeback_control {
 	 */
 	struct swap_iocb **swap_plug;
 
-	/* Target list for splitting a large folio */
-	struct list_head *list;
-
 	/* internal fields used by the ->writepages implementation: */
 	struct folio_batch fbatch;
 	pgoff_t index;
diff --git a/mm/shmem.c b/mm/shmem.c
index 0c5fb4ffa03a..71039b9847c5 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -1540,11 +1540,13 @@ int shmem_unuse(unsigned int type)
 /**
  * shmem_writeout - Write the folio to swap
  * @folio: The folio to write
- * @wbc: How writeback is to be done
+ * @plug: swap plug
+ * @folio_list: list to put back folios on split
  *
  * Move the folio from the page cache to the swap cache.
  */
-int shmem_writeout(struct folio *folio, struct writeback_control *wbc)
+int shmem_writeout(struct folio *folio, struct swap_iocb **plug,
+		struct list_head *folio_list)
 {
 	struct address_space *mapping = folio->mapping;
 	struct inode *inode = mapping->host;
@@ -1554,9 +1556,6 @@ int shmem_writeout(struct folio *folio, struct writeback_control *wbc)
 	int nr_pages;
 	bool split = false;
 
-	if (WARN_ON_ONCE(!wbc->for_reclaim))
-		goto redirty;
-
 	if ((info->flags & VM_LOCKED) || sbinfo->noswap)
 		goto redirty;
 
@@ -1583,7 +1582,7 @@ int shmem_writeout(struct folio *folio, struct writeback_control *wbc)
 try_split:
 		/* Ensure the subpages are still dirty */
 		folio_test_set_dirty(folio);
-		if (split_folio_to_list(folio, wbc->list))
+		if (split_folio_to_list(folio, folio_list))
 			goto redirty;
 		folio_clear_dirty(folio);
 	}
@@ -1636,13 +1635,21 @@ int shmem_writeout(struct folio *folio, struct writeback_control *wbc)
 		list_add(&info->swaplist, &shmem_swaplist);
 
 	if (!folio_alloc_swap(folio, __GFP_HIGH | __GFP_NOMEMALLOC | __GFP_NOWARN)) {
+		struct writeback_control wbc = {
+			.sync_mode	= WB_SYNC_NONE,
+			.nr_to_write	= SWAP_CLUSTER_MAX,
+			.range_start	= 0,
+			.range_end	= LLONG_MAX,
+			.for_reclaim	= 1,
+			.swap_plug	= plug,
+		};
 		shmem_recalc_inode(inode, 0, nr_pages);
 		swap_shmem_alloc(folio->swap, nr_pages);
 		shmem_delete_from_page_cache(folio, swp_to_radix_entry(folio->swap));
 
 		mutex_unlock(&shmem_swaplist_mutex);
 		BUG_ON(folio_mapped(folio));
-		return swap_writeout(folio, wbc);
+		return swap_writeout(folio, &wbc);
 	}
 	if (!info->swapped)
 		list_del_init(&info->swaplist);
@@ -1651,10 +1658,7 @@ int shmem_writeout(struct folio *folio, struct writeback_control *wbc)
 		goto try_split;
 redirty:
 	folio_mark_dirty(folio);
-	if (wbc->for_reclaim)
-		return AOP_WRITEPAGE_ACTIVATE;	/* Return with folio locked */
-	folio_unlock(folio);
-	return 0;
+	return AOP_WRITEPAGE_ACTIVATE;	/* Return with folio locked */
 }
 EXPORT_SYMBOL_GPL(shmem_writeout);
 
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 3af99b0978b5..d56a4e8a1ed9 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -669,15 +669,13 @@ static pageout_t writeout(struct folio *folio, struct address_space *mapping,
 
 	/*
 	 * The large shmem folio can be split if CONFIG_THP_SWAP is not enabled
-	 * or we failed to allocate contiguous swap entries.
+	 * or we failed to allocate contiguous swap entries, in which case
+	 * the split out folios get added back to folio_list.
 	 */
-	if (shmem_mapping(mapping)) {
-		if (folio_test_large(folio))
-			wbc.list = folio_list;
-		res = shmem_writeout(folio, &wbc);
-	} else {
+	if (shmem_mapping(mapping))
+		res = shmem_writeout(folio, plug, folio_list);
+	else
 		res = swap_writeout(folio, &wbc);
-	}
 
 	if (res < 0)
 		handle_write_error(mapping, folio, res);
-- 
2.47.2

Re: [PATCH 2/6] mm: stop passing a writeback_control structure to shmem_writeout

[Baolin Wang]

On 2025/6/10 13:49, Christoph Hellwig wrote:
> shmem_writeout only needs the swap_iocb cookie and the split folio
> list.  Pass those explicitly and remove the now unused list member
> from struct writeback_control.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Thanks for the cleaning up, and I did a quick shmem swap-out/in test, 
and no issues found.

Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Tested-by: Baolin Wang <baolin.wang@linux.alibaba.com>

[PATCH 3/6] mm: tidy up swap_writeout

[Christoph Hellwig]

Use a goto label to consolidate the unlock folio and return pattern
and don't bother with an else after a return / goto.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 mm/page_io.c | 36 ++++++++++++++++++------------------
 1 file changed, 18 insertions(+), 18 deletions(-)

diff --git a/mm/page_io.c b/mm/page_io.c
index f7716b6569fa..c420b0aa0f22 100644
--- a/mm/page_io.c
+++ b/mm/page_io.c
@@ -239,12 +239,11 @@ static void swap_zeromap_folio_clear(struct folio *folio)
  */
 int swap_writeout(struct folio *folio, struct writeback_control *wbc)
 {
-	int ret;
+	int ret = 0;
+
+	if (folio_free_swap(folio))
+		goto out_unlock;
 
-	if (folio_free_swap(folio)) {
-		folio_unlock(folio);
-		return 0;
-	}
 	/*
 	 * Arch code may have to preserve more data than just the page
 	 * contents, e.g. memory tags.
@@ -252,8 +251,7 @@ int swap_writeout(struct folio *folio, struct writeback_control *wbc)
 	ret = arch_prepare_to_swap(folio);
 	if (ret) {
 		folio_mark_dirty(folio);
-		folio_unlock(folio);
-		return ret;
+		goto out_unlock;
 	}
 
 	/*
@@ -264,20 +262,19 @@ int swap_writeout(struct folio *folio, struct writeback_control *wbc)
 	 */
 	if (is_folio_zero_filled(folio)) {
 		swap_zeromap_folio_set(folio);
-		folio_unlock(folio);
-		return 0;
-	} else {
-		/*
-		 * Clear bits this folio occupies in the zeromap to prevent
-		 * zero data being read in from any previous zero writes that
-		 * occupied the same swap entries.
-		 */
-		swap_zeromap_folio_clear(folio);
+		goto out_unlock;
 	}
+
+	/*
+	 * Clear bits this folio occupies in the zeromap to prevent zero data
+	 * being read in from any previous zero writes that occupied the same
+	 * swap entries.
+	 */
+	swap_zeromap_folio_clear(folio);
+
 	if (zswap_store(folio)) {
 		count_mthp_stat(folio_order(folio), MTHP_STAT_ZSWPOUT);
-		folio_unlock(folio);
-		return 0;
+		goto out_unlock;
 	}
 	if (!mem_cgroup_zswap_writeback_enabled(folio_memcg(folio))) {
 		folio_mark_dirty(folio);
@@ -286,6 +283,9 @@ int swap_writeout(struct folio *folio, struct writeback_control *wbc)
 
 	__swap_writepage(folio, wbc);
 	return 0;
+out_unlock:
+	folio_unlock(folio);
+	return ret;
 }
 
 static inline void count_swpout_vm_event(struct folio *folio)
-- 
2.47.2

[PATCH 4/6] mm: stop passing a writeback_control structure to __swap_writepage

[Christoph Hellwig]

__swap_writepage only needs the swap_iocb cookie from the
writeback_control structure, so pass it explicitly and remove the
now unused swap_iocb member from struct writeback_control.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 mm/page_io.c | 33 ++++++++++++++-------------------
 mm/swap.h    |  2 +-
 mm/zswap.c   |  5 +----
 3 files changed, 16 insertions(+), 24 deletions(-)

diff --git a/mm/page_io.c b/mm/page_io.c
index c420b0aa0f22..fb52bedcc966 100644
--- a/mm/page_io.c
+++ b/mm/page_io.c
@@ -281,7 +281,7 @@ int swap_writeout(struct folio *folio, struct writeback_control *wbc)
 		return AOP_WRITEPAGE_ACTIVATE;
 	}
 
-	__swap_writepage(folio, wbc);
+	__swap_writepage(folio, wbc->swap_plug);
 	return 0;
 out_unlock:
 	folio_unlock(folio);
@@ -371,9 +371,9 @@ static void sio_write_complete(struct kiocb *iocb, long ret)
 	mempool_free(sio, sio_pool);
 }
 
-static void swap_writepage_fs(struct folio *folio, struct writeback_control *wbc)
+static void swap_writepage_fs(struct folio *folio, struct swap_iocb **swap_plug)
 {
-	struct swap_iocb *sio = NULL;
+	struct swap_iocb *sio = swap_plug ? *swap_plug : NULL;
 	struct swap_info_struct *sis = swp_swap_info(folio->swap);
 	struct file *swap_file = sis->swap_file;
 	loff_t pos = swap_dev_pos(folio->swap);
@@ -381,8 +381,6 @@ static void swap_writepage_fs(struct folio *folio, struct writeback_control *wbc
 	count_swpout_vm_event(folio);
 	folio_start_writeback(folio);
 	folio_unlock(folio);
-	if (wbc->swap_plug)
-		sio = *wbc->swap_plug;
 	if (sio) {
 		if (sio->iocb.ki_filp != swap_file ||
 		    sio->iocb.ki_pos + sio->len != pos) {
@@ -401,22 +399,21 @@ static void swap_writepage_fs(struct folio *folio, struct writeback_control *wbc
 	bvec_set_folio(&sio->bvec[sio->pages], folio, folio_size(folio), 0);
 	sio->len += folio_size(folio);
 	sio->pages += 1;
-	if (sio->pages == ARRAY_SIZE(sio->bvec) || !wbc->swap_plug) {
+	if (sio->pages == ARRAY_SIZE(sio->bvec) || !swap_plug) {
 		swap_write_unplug(sio);
 		sio = NULL;
 	}
-	if (wbc->swap_plug)
-		*wbc->swap_plug = sio;
+	if (swap_plug)
+		*swap_plug = sio;
 }
 
 static void swap_writepage_bdev_sync(struct folio *folio,
-		struct writeback_control *wbc, struct swap_info_struct *sis)
+		struct swap_info_struct *sis)
 {
 	struct bio_vec bv;
 	struct bio bio;
 
-	bio_init(&bio, sis->bdev, &bv, 1,
-		 REQ_OP_WRITE | REQ_SWAP | wbc_to_write_flags(wbc));
+	bio_init(&bio, sis->bdev, &bv, 1, REQ_OP_WRITE | REQ_SWAP);
 	bio.bi_iter.bi_sector = swap_folio_sector(folio);
 	bio_add_folio_nofail(&bio, folio, folio_size(folio), 0);
 
@@ -431,13 +428,11 @@ static void swap_writepage_bdev_sync(struct folio *folio,
 }
 
 static void swap_writepage_bdev_async(struct folio *folio,
-		struct writeback_control *wbc, struct swap_info_struct *sis)
+		struct swap_info_struct *sis)
 {
 	struct bio *bio;
 
-	bio = bio_alloc(sis->bdev, 1,
-			REQ_OP_WRITE | REQ_SWAP | wbc_to_write_flags(wbc),
-			GFP_NOIO);
+	bio = bio_alloc(sis->bdev, 1, REQ_OP_WRITE | REQ_SWAP, GFP_NOIO);
 	bio->bi_iter.bi_sector = swap_folio_sector(folio);
 	bio->bi_end_io = end_swap_bio_write;
 	bio_add_folio_nofail(bio, folio, folio_size(folio), 0);
@@ -449,7 +444,7 @@ static void swap_writepage_bdev_async(struct folio *folio,
 	submit_bio(bio);
 }
 
-void __swap_writepage(struct folio *folio, struct writeback_control *wbc)
+void __swap_writepage(struct folio *folio, struct swap_iocb **swap_plug)
 {
 	struct swap_info_struct *sis = swp_swap_info(folio->swap);
 
@@ -460,16 +455,16 @@ void __swap_writepage(struct folio *folio, struct writeback_control *wbc)
 	 * is safe.
 	 */
 	if (data_race(sis->flags & SWP_FS_OPS))
-		swap_writepage_fs(folio, wbc);
+		swap_writepage_fs(folio, swap_plug);
 	/*
 	 * ->flags can be updated non-atomicially (scan_swap_map_slots),
 	 * but that will never affect SWP_SYNCHRONOUS_IO, so the data_race
 	 * is safe.
 	 */
 	else if (data_race(sis->flags & SWP_SYNCHRONOUS_IO))
-		swap_writepage_bdev_sync(folio, wbc, sis);
+		swap_writepage_bdev_sync(folio, sis);
 	else
-		swap_writepage_bdev_async(folio, wbc, sis);
+		swap_writepage_bdev_async(folio, sis);
 }
 
 void swap_write_unplug(struct swap_iocb *sio)
diff --git a/mm/swap.h b/mm/swap.h
index 2269eb9df0af..045415e2eb4f 100644
--- a/mm/swap.h
+++ b/mm/swap.h
@@ -21,7 +21,7 @@ static inline void swap_read_unplug(struct swap_iocb *plug)
 }
 void swap_write_unplug(struct swap_iocb *sio);
 int swap_writeout(struct folio *folio, struct writeback_control *wbc);
-void __swap_writepage(struct folio *folio, struct writeback_control *wbc);
+void __swap_writepage(struct folio *folio, struct swap_iocb **swap_plug);
 
 /* linux/mm/swap_state.c */
 /* One swap address space for each 64M swap space */
diff --git a/mm/zswap.c b/mm/zswap.c
index 455e9425c5f5..3c0fd8a13718 100644
--- a/mm/zswap.c
+++ b/mm/zswap.c
@@ -1070,9 +1070,6 @@ static int zswap_writeback_entry(struct zswap_entry *entry,
 	struct mempolicy *mpol;
 	bool folio_was_allocated;
 	struct swap_info_struct *si;
-	struct writeback_control wbc = {
-		.sync_mode = WB_SYNC_NONE,
-	};
 	int ret = 0;
 
 	/* try to allocate swap cache folio */
@@ -1134,7 +1131,7 @@ static int zswap_writeback_entry(struct zswap_entry *entry,
 	folio_set_reclaim(folio);
 
 	/* start writeback */
-	__swap_writepage(folio, &wbc);
+	__swap_writepage(folio, NULL);
 
 out:
 	if (ret && ret != -EEXIST) {
-- 
2.47.2

Re: [PATCH 4/6] mm: stop passing a writeback_control structure to __swap_writepage

[Nhat Pham]

On Mon, Jun 9, 2025 at 10:50 PM Christoph Hellwig <hch@lst.de> wrote:
>
> __swap_writepage only needs the swap_iocb cookie from the
> writeback_control structure, so pass it explicitly and remove the
> now unused swap_iocb member from struct writeback_control.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Acked-by: Nhat Pham <nphamcs@gmail.com>

Thanks!

[PATCH 5/6] mm: stop passing a writeback_control structure to swap_writeout

[Christoph Hellwig]

swap_writeout only needs the swap_iocb cookie from the writeback_control
structure, so pass it explicitly.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 include/linux/writeback.h |  7 -------
 mm/page_io.c              |  4 ++--
 mm/shmem.c                | 10 +---------
 mm/swap.h                 |  7 +++++--
 mm/vmscan.c               | 10 +---------
 5 files changed, 9 insertions(+), 29 deletions(-)

diff --git a/include/linux/writeback.h b/include/linux/writeback.h
index 82f217970092..9e960f2faf79 100644
--- a/include/linux/writeback.h
+++ b/include/linux/writeback.h
@@ -72,13 +72,6 @@ struct writeback_control {
 	 */
 	unsigned no_cgroup_owner:1;
 
-	/* To enable batching of swap writes to non-block-device backends,
-	 * "plug" can be set point to a 'struct swap_iocb *'.  When all swap
-	 * writes have been submitted, if with swap_iocb is not NULL,
-	 * swap_write_unplug() should be called.
-	 */
-	struct swap_iocb **swap_plug;
-
 	/* internal fields used by the ->writepages implementation: */
 	struct folio_batch fbatch;
 	pgoff_t index;
diff --git a/mm/page_io.c b/mm/page_io.c
index fb52bedcc966..a2056a5ecb13 100644
--- a/mm/page_io.c
+++ b/mm/page_io.c
@@ -237,7 +237,7 @@ static void swap_zeromap_folio_clear(struct folio *folio)
  * We may have stale swap cache pages in memory: notice
  * them here and get rid of the unnecessary final write.
  */
-int swap_writeout(struct folio *folio, struct writeback_control *wbc)
+int swap_writeout(struct folio *folio, struct swap_iocb **swap_plug)
 {
 	int ret = 0;
 
@@ -281,7 +281,7 @@ int swap_writeout(struct folio *folio, struct writeback_control *wbc)
 		return AOP_WRITEPAGE_ACTIVATE;
 	}
 
-	__swap_writepage(folio, wbc->swap_plug);
+	__swap_writepage(folio, swap_plug);
 	return 0;
 out_unlock:
 	folio_unlock(folio);
diff --git a/mm/shmem.c b/mm/shmem.c
index 71039b9847c5..b649f181e2e1 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -1635,21 +1635,13 @@ int shmem_writeout(struct folio *folio, struct swap_iocb **plug,
 		list_add(&info->swaplist, &shmem_swaplist);
 
 	if (!folio_alloc_swap(folio, __GFP_HIGH | __GFP_NOMEMALLOC | __GFP_NOWARN)) {
-		struct writeback_control wbc = {
-			.sync_mode	= WB_SYNC_NONE,
-			.nr_to_write	= SWAP_CLUSTER_MAX,
-			.range_start	= 0,
-			.range_end	= LLONG_MAX,
-			.for_reclaim	= 1,
-			.swap_plug	= plug,
-		};
 		shmem_recalc_inode(inode, 0, nr_pages);
 		swap_shmem_alloc(folio->swap, nr_pages);
 		shmem_delete_from_page_cache(folio, swp_to_radix_entry(folio->swap));
 
 		mutex_unlock(&shmem_swaplist_mutex);
 		BUG_ON(folio_mapped(folio));
-		return swap_writeout(folio, &wbc);
+		return swap_writeout(folio, plug);
 	}
 	if (!info->swapped)
 		list_del_init(&info->swaplist);
diff --git a/mm/swap.h b/mm/swap.h
index 045415e2eb4f..e87a0f19a0ee 100644
--- a/mm/swap.h
+++ b/mm/swap.h
@@ -3,6 +3,8 @@
 #define _MM_SWAP_H
 
 struct mempolicy;
+struct swap_iocb;
+
 extern int page_cluster;
 
 #ifdef CONFIG_SWAP
@@ -20,7 +22,7 @@ static inline void swap_read_unplug(struct swap_iocb *plug)
 		__swap_read_unplug(plug);
 }
 void swap_write_unplug(struct swap_iocb *sio);
-int swap_writeout(struct folio *folio, struct writeback_control *wbc);
+int swap_writeout(struct folio *folio, struct swap_iocb **swap_plug);
 void __swap_writepage(struct folio *folio, struct swap_iocb **swap_plug);
 
 /* linux/mm/swap_state.c */
@@ -141,7 +143,8 @@ static inline struct folio *swapin_readahead(swp_entry_t swp, gfp_t gfp_mask,
 	return NULL;
 }
 
-static inline int swap_writeout(struct folio *f, struct writeback_control *wbc)
+static inline int swap_writeout(struct folio *folio,
+		struct swap_iocb **swap_plug)
 {
 	return 0;
 }
diff --git a/mm/vmscan.c b/mm/vmscan.c
index d56a4e8a1ed9..ae16e57bf28e 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -655,14 +655,6 @@ typedef enum {
 static pageout_t writeout(struct folio *folio, struct address_space *mapping,
 		struct swap_iocb **plug, struct list_head *folio_list)
 {
-	struct writeback_control wbc = {
-		.sync_mode = WB_SYNC_NONE,
-		.nr_to_write = SWAP_CLUSTER_MAX,
-		.range_start = 0,
-		.range_end = LLONG_MAX,
-		.for_reclaim = 1,
-		.swap_plug = plug,
-	};
 	int res;
 
 	folio_set_reclaim(folio);
@@ -675,7 +667,7 @@ static pageout_t writeout(struct folio *folio, struct address_space *mapping,
 	if (shmem_mapping(mapping))
 		res = shmem_writeout(folio, plug, folio_list);
 	else
-		res = swap_writeout(folio, &wbc);
+		res = swap_writeout(folio, plug);
 
 	if (res < 0)
 		handle_write_error(mapping, folio, res);
-- 
2.47.2

[PATCH 6/6] mm: remove the for_reclaim field from struct writeback_control

[Christoph Hellwig]

This field is now only set to one in the i915 gem code that only
calls writeback_iter on it, which ignores the flag.  All other
checks are thuse dead code and the field can be removed.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/gpu/drm/i915/gem/i915_gem_shmem.c |  1 -
 fs/fuse/file.c                            | 11 -----------
 fs/nfs/write.c                            |  2 +-
 include/linux/writeback.h                 |  1 -
 include/trace/events/btrfs.h              |  7 ++-----
 include/trace/events/writeback.h          |  8 ++------
 6 files changed, 5 insertions(+), 25 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
index 24d8daa4fdb3..f263615f6ece 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
@@ -302,7 +302,6 @@ void __shmem_writeback(size_t size, struct address_space *mapping)
 		.nr_to_write = SWAP_CLUSTER_MAX,
 		.range_start = 0,
 		.range_end = LLONG_MAX,
-		.for_reclaim = 1,
 	};
 	struct folio *folio = NULL;
 	int error = 0;
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index f102afc03359..2ad5665bf81d 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -1928,17 +1928,6 @@ int fuse_write_inode(struct inode *inode, struct writeback_control *wbc)
 	struct fuse_file *ff;
 	int err;
 
-	/*
-	 * Inode is always written before the last reference is dropped and
-	 * hence this should not be reached from reclaim.
-	 *
-	 * Writing back the inode from reclaim can deadlock if the request
-	 * processing itself needs an allocation.  Allocations triggering
-	 * reclaim while serving a request can't be prevented, because it can
-	 * involve any number of unrelated userspace processes.
-	 */
-	WARN_ON(wbc->for_reclaim);
-
 	ff = __fuse_write_file_get(fi);
 	err = fuse_flush_times(inode, ff);
 	if (ff)
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 374fc6b34c79..cf1d720b8251 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -720,7 +720,7 @@ int nfs_writepages(struct address_space *mapping, struct writeback_control *wbc)
 	nfs_inc_stats(inode, NFSIOS_VFSWRITEPAGES);
 
 	if (!(mntflags & NFS_MOUNT_WRITE_EAGER) || wbc->for_kupdate ||
-	    wbc->for_background || wbc->for_sync || wbc->for_reclaim) {
+	    wbc->for_background || wbc->for_sync) {
 		ioc = nfs_io_completion_alloc(GFP_KERNEL);
 		if (ioc)
 			nfs_io_completion_init(ioc, nfs_io_completion_commit,
diff --git a/include/linux/writeback.h b/include/linux/writeback.h
index 9e960f2faf79..a2848d731a46 100644
--- a/include/linux/writeback.h
+++ b/include/linux/writeback.h
@@ -59,7 +59,6 @@ struct writeback_control {
 	unsigned for_kupdate:1;		/* A kupdate writeback */
 	unsigned for_background:1;	/* A background writeback */
 	unsigned tagged_writepages:1;	/* tag-and-write to avoid livelock */
-	unsigned for_reclaim:1;		/* Invoked from the page allocator */
 	unsigned range_cyclic:1;	/* range_start is cyclic */
 	unsigned for_sync:1;		/* sync(2) WB_SYNC_ALL writeback */
 	unsigned unpinned_netfs_wb:1;	/* Cleared I_PINNING_NETFS_WB */
diff --git a/include/trace/events/btrfs.h b/include/trace/events/btrfs.h
index bebc252db865..0adc40f5e72b 100644
--- a/include/trace/events/btrfs.h
+++ b/include/trace/events/btrfs.h
@@ -686,7 +686,6 @@ DECLARE_EVENT_CLASS(btrfs__writepage,
 		__field(	loff_t, range_start		)
 		__field(	loff_t, range_end		)
 		__field(	char,   for_kupdate		)
-		__field(	char,   for_reclaim		)
 		__field(	char,   range_cyclic		)
 		__field(	unsigned long,  writeback_index	)
 		__field(	u64,    root_objectid		)
@@ -700,7 +699,6 @@ DECLARE_EVENT_CLASS(btrfs__writepage,
 		__entry->range_start	= wbc->range_start;
 		__entry->range_end	= wbc->range_end;
 		__entry->for_kupdate	= wbc->for_kupdate;
-		__entry->for_reclaim	= wbc->for_reclaim;
 		__entry->range_cyclic	= wbc->range_cyclic;
 		__entry->writeback_index = inode->i_mapping->writeback_index;
 		__entry->root_objectid	= btrfs_root_id(BTRFS_I(inode)->root);
@@ -709,13 +707,12 @@ DECLARE_EVENT_CLASS(btrfs__writepage,
 	TP_printk_btrfs("root=%llu(%s) ino=%llu page_index=%lu "
 		  "nr_to_write=%ld pages_skipped=%ld range_start=%llu "
 		  "range_end=%llu for_kupdate=%d "
-		  "for_reclaim=%d range_cyclic=%d writeback_index=%lu",
+		  "range_cyclic=%d writeback_index=%lu",
 		  show_root_type(__entry->root_objectid),
 		  __entry->ino, __entry->index,
 		  __entry->nr_to_write, __entry->pages_skipped,
 		  __entry->range_start, __entry->range_end,
-		  __entry->for_kupdate,
-		  __entry->for_reclaim, __entry->range_cyclic,
+		  __entry->for_kupdate, __entry->range_cyclic,
 		  __entry->writeback_index)
 );
 
diff --git a/include/trace/events/writeback.h b/include/trace/events/writeback.h
index 0ff388131fc9..1e23919c0da9 100644
--- a/include/trace/events/writeback.h
+++ b/include/trace/events/writeback.h
@@ -459,7 +459,6 @@ DECLARE_EVENT_CLASS(wbc_class,
 		__field(int, sync_mode)
 		__field(int, for_kupdate)
 		__field(int, for_background)
-		__field(int, for_reclaim)
 		__field(int, range_cyclic)
 		__field(long, range_start)
 		__field(long, range_end)
@@ -473,23 +472,20 @@ DECLARE_EVENT_CLASS(wbc_class,
 		__entry->sync_mode	= wbc->sync_mode;
 		__entry->for_kupdate	= wbc->for_kupdate;
 		__entry->for_background	= wbc->for_background;
-		__entry->for_reclaim	= wbc->for_reclaim;
 		__entry->range_cyclic	= wbc->range_cyclic;
 		__entry->range_start	= (long)wbc->range_start;
 		__entry->range_end	= (long)wbc->range_end;
 		__entry->cgroup_ino	= __trace_wbc_assign_cgroup(wbc);
 	),
 
-	TP_printk("bdi %s: towrt=%ld skip=%ld mode=%d kupd=%d "
-		"bgrd=%d reclm=%d cyclic=%d "
-		"start=0x%lx end=0x%lx cgroup_ino=%lu",
+	TP_printk("bdi %s: towrt=%ld skip=%ld mode=%d kupd=%d bgrd=%d "
+		"cyclic=%d start=0x%lx end=0x%lx cgroup_ino=%lu",
 		__entry->name,
 		__entry->nr_to_write,
 		__entry->pages_skipped,
 		__entry->sync_mode,
 		__entry->for_kupdate,
 		__entry->for_background,
-		__entry->for_reclaim,
 		__entry->range_cyclic,
 		__entry->range_start,
 		__entry->range_end,
-- 
2.47.2