The patch titled
Subject: mm: drop swap_dirty_folio
has been added to the -mm tree. Its filename is
mm-drop-swap_dirty_folio.patch
This patch should soon appear at
https://ozlabs.org/~akpm/mmots/broken-out/mm-drop-swap_dirty_folio.patch
and later at
https://ozlabs.org/~akpm/mmotm/broken-out/mm-drop-swap_dirty_folio.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***
The -mm tree is included into linux-next and is updated
there every 3-4 working days
------------------------------------------------------
From: NeilBrown <neilb@xxxxxxx>
Subject: mm: drop swap_dirty_folio
folios that are written to swap are owned by the MM subsystem - not any
filesystem.
When such a folio is passed to a filesystem to be written out to a
swap-file, the filesystem handles the data, but the folio itself does not
belong to the filesystem. So calling the filesystem's ->dirty_folio()
address_space operation makes no sense. This is for folios in the given
address space, and a folio to be written to swap does not exist in the
given address space.
So drop swap_dirty_folio() which calls the address-space's
->dirty_folio(), and always use noop_dirty_folio(), which is appropriate
for folios being swapped out.
Link: https://lkml.kernel.org/r/164859778123.29473.6900942583784889976.stgit@noble.brown
Signed-off-by: NeilBrown <neilb@xxxxxxx>
Reviewed-by: Christoph Hellwig <hch@xxxxxx>
Cc: David Howells <dhowells@xxxxxxxxxx>
Cc: Hugh Dickins <hughd@xxxxxxxxxx>
Cc: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx>
Cc: Trond Myklebust <trond.myklebust@xxxxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
include/linux/swap.h | 1 -
mm/page_io.c | 17 -----------------
mm/swap_state.c | 2 +-
3 files changed, 1 insertion(+), 19 deletions(-)
--- a/include/linux/swap.h~mm-drop-swap_dirty_folio
+++ a/include/linux/swap.h
@@ -420,7 +420,6 @@ extern void kswapd_stop(int nid);
#ifdef CONFIG_SWAP
-bool swap_dirty_folio(struct address_space *mapping, struct folio *folio);
int add_swap_extent(struct swap_info_struct *sis, unsigned long start_page,
unsigned long nr_pages, sector_t start_block);
int generic_swapfile_activate(struct swap_info_struct *, struct file *,
--- a/mm/page_io.c~mm-drop-swap_dirty_folio
+++ a/mm/page_io.c
@@ -385,20 +385,3 @@ out:
delayacct_swapin_end();
return ret;
}
-
-bool swap_dirty_folio(struct address_space *mapping, struct folio *folio)
-{
- struct swap_info_struct *sis = swp_swap_info(folio_swap_entry(folio));
-
- if (data_race(sis->flags & SWP_FS_OPS)) {
- const struct address_space_operations *aops;
-
- mapping = sis->swap_file->f_mapping;
- aops = mapping->a_ops;
-
- VM_BUG_ON_FOLIO(!folio_test_swapcache(folio), folio);
- return aops->dirty_folio(mapping, folio);
- } else {
- return noop_dirty_folio(mapping, folio);
- }
-}
--- a/mm/swap_state.c~mm-drop-swap_dirty_folio
+++ a/mm/swap_state.c
@@ -31,7 +31,7 @@
*/
static const struct address_space_operations swap_aops = {
.writepage = swap_writepage,
- .dirty_folio = swap_dirty_folio,
+ .dirty_folio = noop_dirty_folio,
#ifdef CONFIG_MIGRATION
.migratepage = migrate_page,
#endif
_
Patches currently in -mm which might be from neilb@xxxxxxx are
mm-create-new-mm-swaph-header-file.patch
mm-drop-swap_dirty_folio.patch
mm-move-responsibility-for-setting-swp_fs_ops-to-swap_activate.patch
mm-reclaim-mustnt-enter-fs-for-swp_fs_ops-swap-space.patch
mm-introduce-swap_rw-and-use-it-for-reads-from-swp_fs_ops-swap-space.patch
mm-perform-async-writes-to-swp_fs_ops-swap-space-using-swap_rw.patch
doc-update-documentation-for-swap_activate-and-swap_rw.patch
mm-submit-multipage-reads-for-swp_fs_ops-swap-space.patch
mm-submit-multipage-write-for-swp_fs_ops-swap-space.patch
vfs-add-fmode_can_odirect-file-flag.patch
mm-discard-__gfp_atomic.patch
