LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* [PATCH v3] btrfs: implement migratepage callback
@ 2020-03-05  0:57 Roman Gushchin
  2020-03-11  0:40 ` David Sterba
  0 siblings, 1 reply; 3+ messages in thread
From: Roman Gushchin @ 2020-03-05  0:57 UTC (permalink / raw)
  To: Josef Bacik, Chris Mason, David Sterba
  Cc: linux-btrfs, linux-kernel, kernel-team, Rik van Riel, Roman Gushchin

Currently btrfs doesn't provide a migratepage callback. It means that
fallback_migrate_page()	is used to migrate btrfs pages.

fallback_migrate_page() cannot move dirty pages, instead it tries to
flush them (in sync mode) or just fails (in async mode).

In the sync mode pages which are scheduled to be processed by
btrfs_writepage_fixup_worker() can't be effectively flushed by the
migration code, because there is no established way to wait for the
completion of the delayed work.

It all leads to page migration failures.

To fix it the patch implements a btrs-specific migratepage callback,
which is similar to iomap_migrate_page() used by some other fs, except
it does take care of the PagePrivate2 flag which is used for data
ordering purposes.

v3: fixed the build issue once again
v2: fixed the build issue found by the kbuild test robot <lkp@intel.com>

Signed-off-by: Roman Gushchin <guro@fb.com>
Reviewed-by: Chris Mason <clm@fb.com>
---
 fs/btrfs/inode.c | 37 +++++++++++++++++++++++++++++++++++++
 1 file changed, 37 insertions(+)

diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 7735ce6127c3..6ed6df4afe5a 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -28,6 +28,7 @@
 #include <linux/magic.h>
 #include <linux/iversion.h>
 #include <linux/swap.h>
+#include <linux/migrate.h>
 #include <linux/sched/mm.h>
 #include <asm/unaligned.h>
 #include "misc.h"
@@ -8323,6 +8324,39 @@ static int btrfs_releasepage(struct page *page, gfp_t gfp_flags)
 	return __btrfs_releasepage(page, gfp_flags);
 }
 
+#ifdef CONFIG_MIGRATION
+static int btrfs_migratepage(struct address_space *mapping,
+			     struct page *newpage, struct page *page,
+			     enum migrate_mode mode)
+{
+	int ret;
+
+	ret = migrate_page_move_mapping(mapping, newpage, page, 0);
+	if (ret != MIGRATEPAGE_SUCCESS)
+		return ret;
+
+	if (page_has_private(page)) {
+		ClearPagePrivate(page);
+		get_page(newpage);
+		set_page_private(newpage, page_private(page));
+		set_page_private(page, 0);
+		put_page(page);
+		SetPagePrivate(newpage);
+	}
+
+	if (PagePrivate2(page)) {
+		ClearPagePrivate2(page);
+		SetPagePrivate2(newpage);
+	}
+
+	if (mode != MIGRATE_SYNC_NO_COPY)
+		migrate_page_copy(newpage, page);
+	else
+		migrate_page_states(newpage, page);
+	return MIGRATEPAGE_SUCCESS;
+}
+#endif
+
 static void btrfs_invalidatepage(struct page *page, unsigned int offset,
 				 unsigned int length)
 {
@@ -10525,6 +10559,9 @@ static const struct address_space_operations btrfs_aops = {
 	.direct_IO	= btrfs_direct_IO,
 	.invalidatepage = btrfs_invalidatepage,
 	.releasepage	= btrfs_releasepage,
+#ifdef CONFIG_MIGRATION
+	.migratepage	= btrfs_migratepage,
+#endif
 	.set_page_dirty	= btrfs_set_page_dirty,
 	.error_remove_page = generic_error_remove_page,
 	.swap_activate	= btrfs_swap_activate,
-- 
2.24.1


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH v3] btrfs: implement migratepage callback
  2020-03-05  0:57 [PATCH v3] btrfs: implement migratepage callback Roman Gushchin
@ 2020-03-11  0:40 ` David Sterba
  2020-03-11  1:30   ` Roman Gushchin
  0 siblings, 1 reply; 3+ messages in thread
From: David Sterba @ 2020-03-11  0:40 UTC (permalink / raw)
  To: Roman Gushchin
  Cc: Josef Bacik, Chris Mason, David Sterba, linux-btrfs,
	linux-kernel, kernel-team, Rik van Riel

On Wed, Mar 04, 2020 at 04:57:35PM -0800, Roman Gushchin wrote:
> Currently btrfs doesn't provide a migratepage callback. It means that
> fallback_migrate_page()	is used to migrate btrfs pages.

The callback exists for the metadata pages (btree_migratepage), so I've
added 'for data pages' where appropriate.

> fallback_migrate_page() cannot move dirty pages, instead it tries to
> flush them (in sync mode) or just fails (in async mode).
> 
> In the sync mode pages which are scheduled to be processed by
> btrfs_writepage_fixup_worker() can't be effectively flushed by the
> migration code, because there is no established way to wait for the
> completion of the delayed work.
> 
> It all leads to page migration failures.
> 
> To fix it the patch implements a btrs-specific migratepage callback,
> which is similar to iomap_migrate_page() used by some other fs, except
> it does take care of the PagePrivate2 flag which is used for data
> ordering purposes.
> 
> v3: fixed the build issue once again
> v2: fixed the build issue found by the kbuild test robot <lkp@intel.com>
> 
> Signed-off-by: Roman Gushchin <guro@fb.com>
> Reviewed-by: Chris Mason <clm@fb.com>

Added to devel queue, thanks.

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH v3] btrfs: implement migratepage callback
  2020-03-11  0:40 ` David Sterba
@ 2020-03-11  1:30   ` Roman Gushchin
  0 siblings, 0 replies; 3+ messages in thread
From: Roman Gushchin @ 2020-03-11  1:30 UTC (permalink / raw)
  To: dsterba, Josef Bacik, Chris Mason, David Sterba, linux-btrfs,
	linux-kernel, kernel-team, Rik van Riel

On Wed, Mar 11, 2020 at 01:40:12AM +0100, David Sterba wrote:
> On Wed, Mar 04, 2020 at 04:57:35PM -0800, Roman Gushchin wrote:
> > Currently btrfs doesn't provide a migratepage callback. It means that
> > fallback_migrate_page()	is used to migrate btrfs pages.
> 
> The callback exists for the metadata pages (btree_migratepage), so I've
> added 'for data pages' where appropriate.

Agree.

> 
> > fallback_migrate_page() cannot move dirty pages, instead it tries to
> > flush them (in sync mode) or just fails (in async mode).
> > 
> > In the sync mode pages which are scheduled to be processed by
> > btrfs_writepage_fixup_worker() can't be effectively flushed by the
> > migration code, because there is no established way to wait for the
> > completion of the delayed work.
> > 
> > It all leads to page migration failures.
> > 
> > To fix it the patch implements a btrs-specific migratepage callback,
> > which is similar to iomap_migrate_page() used by some other fs, except
> > it does take care of the PagePrivate2 flag which is used for data
> > ordering purposes.
> > 
> > v3: fixed the build issue once again
> > v2: fixed the build issue found by the kbuild test robot <lkp@intel.com>
> > 
> > Signed-off-by: Roman Gushchin <guro@fb.com>
> > Reviewed-by: Chris Mason <clm@fb.com>
> 
> Added to devel queue, thanks.

Thank you!

Roman

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2020-03-11  1:31 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-03-05  0:57 [PATCH v3] btrfs: implement migratepage callback Roman Gushchin
2020-03-11  0:40 ` David Sterba
2020-03-11  1:30   ` Roman Gushchin

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).