[PATCH v2 1/2] mm/damon/ops-common: putback folios on invalid migrate nid

[PATCH v2 1/2] mm/damon/ops-common: putback folios on invalid migrate nid

[dayou5941]

From: liyouhong <liyouhong@kylinos.cn>

damon_pa_migrate() and damos_va_migrate() isolate folios into a local list
and then call damon_migrate_pages().  When target_nid is invalid (including
the scheme default NUMA_NO_NODE / -1), damon_migrate_pages() returns early
without putting the folios back to the LRU.

Callers then discard the list head while those folios remain isolated with
an extra reference taken by folio_isolate_lru().  The pages stay off the
LRU for as long as the mapping exists (anon active+inactive counts drop
while RSS does not), and the leftover references can pin the pages after
the mapping is gone.

Put the folios back on the invalid-nid path so ignored migration requests
still return them to the LRU.

Fixes: 7e6c3130690a ("mm/damon/ops-common: ignore migration request to invalid nodes")
Cc: <stable@vger.kernel.org>
Assisted-by: Cursor:grok-4.5
Signed-off-by: liyouhong <liyouhong@kylinos.cn>

---
v2:
- Drop the putback helper for easier stable backport
- Open-code putback only on the invalid-nid path
- Keep the original indentation and Cc stable@
- Add a separate cleanup as 2/2 (factor out damon_putback_folio_list)

---
 mm/damon/ops-common.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/mm/damon/ops-common.c b/mm/damon/ops-common.c
index d1842e2b00ef..f5ded45fabd1 100644
--- a/mm/damon/ops-common.c
+++ b/mm/damon/ops-common.c
@@ -394,8 +394,15 @@ unsigned long damon_migrate_pages(struct list_head *folio_list, int target_nid)
 		return nr_migrated;
 
 	if (target_nid < 0 || target_nid >= MAX_NUMNODES ||
-			!node_state(target_nid, N_MEMORY))
+			!node_state(target_nid, N_MEMORY)) {
+		while (!list_empty(folio_list)) {
+			struct folio *folio = lru_to_folio(folio_list);
+
+			list_del(&folio->lru);
+			folio_putback_lru(folio);
+		}
 		return nr_migrated;
+	}
 
 	noreclaim_flag = memalloc_noreclaim_save();
 
-- 
2.25.1

[PATCH v2 2/2] mm/damon/ops-common: factor out damon_putback_folio_list()

[dayou5941]

From: liyouhong <liyouhong@kylinos.cn>

The putback loop is duplicated in damon_migrate_folio_list() and on the
invalid-nid path of damon_migrate_pages().  Factor it into a small helper
for readability.  No functional change.

Signed-off-by: liyouhong <liyouhong@kylinos.cn>
---
 mm/damon/ops-common.c | 25 +++++++++++++------------
 1 file changed, 13 insertions(+), 12 deletions(-)

diff --git a/mm/damon/ops-common.c b/mm/damon/ops-common.c
index f5ded45fabd1..9a1e8aec5444 100644
--- a/mm/damon/ops-common.c
+++ b/mm/damon/ops-common.c
@@ -331,12 +331,22 @@ static unsigned int __damon_migrate_folio_list(
 	return nr_succeeded;
 }
 
+static void damon_putback_folio_list(struct list_head *folio_list)
+{
+	struct folio *folio;
+
+	while (!list_empty(folio_list)) {
+		folio = lru_to_folio(folio_list);
+		list_del(&folio->lru);
+		folio_putback_lru(folio);
+	}
+}
+
 static unsigned int damon_migrate_folio_list(struct list_head *folio_list,
 						struct pglist_data *pgdat,
 						int target_nid)
 {
 	unsigned int nr_migrated = 0;
-	struct folio *folio;
 	LIST_HEAD(ret_folios);
 	LIST_HEAD(migrate_folios);
 
@@ -374,11 +384,7 @@ static unsigned int damon_migrate_folio_list(struct list_head *folio_list,
 
 	list_splice(&ret_folios, folio_list);
 
-	while (!list_empty(folio_list)) {
-		folio = lru_to_folio(folio_list);
-		list_del(&folio->lru);
-		folio_putback_lru(folio);
-	}
+	damon_putback_folio_list(folio_list);
 
 	return nr_migrated;
 }
@@ -395,12 +401,7 @@ unsigned long damon_migrate_pages(struct list_head *folio_list, int target_nid)
 
 	if (target_nid < 0 || target_nid >= MAX_NUMNODES ||
 			!node_state(target_nid, N_MEMORY)) {
-		while (!list_empty(folio_list)) {
-			struct folio *folio = lru_to_folio(folio_list);
-
-			list_del(&folio->lru);
-			folio_putback_lru(folio);
-		}
+		damon_putback_folio_list(folio_list);
 		return nr_migrated;
 	}
 
-- 
2.25.1

Re: [PATCH v2 2/2] mm/damon/ops-common: factor out damon_putback_folio_list()

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] Removal of the `struct folio *folio;` local variable declaration causes a build failure.
--

--- Patch [2]: [PATCH v2 2/2] mm/damon/ops-common: factor out damon_putback_folio_list() ---
commit 44363c3cc00338e3c3445bc5730a2655a4c0fded
Author: liyouhong <liyouhong@kylinos.cn>
mm/damon/ops-common: factor out damon_putback_folio_list()

This patch factors out the duplicated folio putback loop from
damon_migrate_folio_list() and damon_migrate_pages() into a new helper
function, damon_putback_folio_list().

> diff --git a/mm/damon/ops-common.c b/mm/damon/ops-common.c
> index 0981a5e3af172..4b99c31122d6d 100644
> --- a/mm/damon/ops-common.c
> +++ b/mm/damon/ops-common.c
[ ... ]
> @@ -330,12 +330,22 @@ static unsigned int __damon_migrate_folio_list(
[ ... ]
>  static unsigned int damon_migrate_folio_list(struct list_head *folio_list,
>  						struct pglist_data *pgdat,
>  						int target_nid)
>  {
>  	unsigned int nr_migrated = 0;
> -	struct folio *folio;

[Severity: Low]
Does removing this declaration result in a build failure? The folio variable
appears to still be used in damon_migrate_folio_list() during the initial
loop over folio_list:

	while (!list_empty(folio_list)) {
		cond_resched();

		folio = lru_to_folio(folio_list);
		list_del(&folio->lru);
        ...

>  	LIST_HEAD(ret_folios);
>  	LIST_HEAD(migrate_folios);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260724060135.3014676-1-dayou5941@163.com?part=2

Re: [PATCH v2 1/2] mm/damon/ops-common: putback folios on invalid migrate nid

[SJ Park]

'checkpatch.pl --nogit --nogit-fallback' suggests adding below recipients.  I
added them.  Please consider using get_maintainer.pl from next time.

- Joshua Hahn <joshua.hahnjy@gmail.com>
- linux-mm@kvack.org
- linux-kernel@vger.kernel.org


On Fri, 24 Jul 2026 14:01:34 +0800 dayou5941@163.com wrote:

> From: liyouhong <liyouhong@kylinos.cn>
> 
> damon_pa_migrate() and damos_va_migrate() isolate folios into a local list
> and then call damon_migrate_pages().  When target_nid is invalid (including
> the scheme default NUMA_NO_NODE / -1), damon_migrate_pages() returns early
> without putting the folios back to the LRU.
> 
> Callers then discard the list head while those folios remain isolated with
> an extra reference taken by folio_isolate_lru().  The pages stay off the
> LRU for as long as the mapping exists (anon active+inactive counts drop
> while RSS does not), and the leftover references can pin the pages after
> the mapping is gone.
> 
> Put the folios back on the invalid-nid path so ignored migration requests
> still return them to the LRU.

Thank you for finding and fixing this!

> 
> Fixes: 7e6c3130690a ("mm/damon/ops-common: ignore migration request to invalid nodes")
> Cc: <stable@vger.kernel.org>
> Assisted-by: Cursor:grok-4.5
> Signed-off-by: liyouhong <liyouhong@kylinos.cn>

Reviewed-by: SJ Park <sj@kernel.org>


Thanks,
SJ

[...]

Re: [PATCH v2 1/2] mm/damon/ops-common: putback folios on invalid migrate nid

[SJ Park]

On Fri, 24 Jul 2026 07:26:38 -0700 SJ Park <sj@kernel.org> wrote:

> 
> 'checkpatch.pl --nogit --nogit-fallback' suggests adding below recipients.  I

I mean, get_maintainer.pl.  Sorry if it confused you.


THanks,
SJ

[...]

Re: [PATCH v2 2/2] mm/damon/ops-common: factor out damon_putback_folio_list()

[SJ Park]

'get_maintainer.pl --nogit --nogit-fallback' suggests adding below recipients.
I added them.  Please consier using get_maintainer.pl from the next time.

- linux-mm@kvack.org
- linux-kernel@vger.kernel.org

On Fri, 24 Jul 2026 14:01:35 +0800 dayou5941@163.com wrote:

> From: liyouhong <liyouhong@kylinos.cn>
> 
> The putback loop is duplicated in damon_migrate_folio_list() and on the
> invalid-nid path of damon_migrate_pages().  Factor it into a small helper
> for readability.  No functional change.

This is not a hotfix.  I'd suggest sending this separately, not together with
the first patch of this series.  Sending hotfix together with non-hotfix when
they don't really need to be applied together only makes it complicated.

> 
> Signed-off-by: liyouhong <liyouhong@kylinos.cn>
> ---
>  mm/damon/ops-common.c | 25 +++++++++++++------------
>  1 file changed, 13 insertions(+), 12 deletions(-)
> 
> diff --git a/mm/damon/ops-common.c b/mm/damon/ops-common.c
> index f5ded45fabd1..9a1e8aec5444 100644
> --- a/mm/damon/ops-common.c
> +++ b/mm/damon/ops-common.c
> @@ -331,12 +331,22 @@ static unsigned int __damon_migrate_folio_list(
>  	return nr_succeeded;
>  }
>  
> +static void damon_putback_folio_list(struct list_head *folio_list)
> +{
> +	struct folio *folio;
> +
> +	while (!list_empty(folio_list)) {
> +		folio = lru_to_folio(folio_list);
> +		list_del(&folio->lru);
> +		folio_putback_lru(folio);
> +	}
> +}
> +
>  static unsigned int damon_migrate_folio_list(struct list_head *folio_list,
>  						struct pglist_data *pgdat,
>  						int target_nid)
>  {
>  	unsigned int nr_migrated = 0;
> -	struct folio *folio;
>  	LIST_HEAD(ret_folios);
>  	LIST_HEAD(migrate_folios);
>  
> @@ -374,11 +384,7 @@ static unsigned int damon_migrate_folio_list(struct list_head *folio_list,
>  
>  	list_splice(&ret_folios, folio_list);
>  
> -	while (!list_empty(folio_list)) {
> -		folio = lru_to_folio(folio_list);
> -		list_del(&folio->lru);
> -		folio_putback_lru(folio);
> -	}
> +	damon_putback_folio_list(folio_list);
>  
>  	return nr_migrated;
>  }
> @@ -395,12 +401,7 @@ unsigned long damon_migrate_pages(struct list_head *folio_list, int target_nid)
>  
>  	if (target_nid < 0 || target_nid >= MAX_NUMNODES ||
>  			!node_state(target_nid, N_MEMORY)) {
> -		while (!list_empty(folio_list)) {
> -			struct folio *folio = lru_to_folio(folio_list);
> -
> -			list_del(&folio->lru);
> -			folio_putback_lru(folio);
> -		}
> +		damon_putback_folio_list(folio_list);
>  		return nr_migrated;
>  	}

Looks better.  But, how about further simplifying it by moving the folios
putback from damon_migrate_pages(), and doing that from damon_migrate_pages()?
damon_migrate_pages() would do the putback always before returning, and the
taregt_nid path will 'goto' the path.  E.g.,

--- a/mm/damon/ops-common.c
+++ b/mm/damon/ops-common.c
@@ -391,15 +391,8 @@ unsigned long damon_migrate_pages(struct list_head *folio_list, int target_nid)
                return nr_migrated;

        if (target_nid < 0 || target_nid >= MAX_NUMNODES ||
-                       !node_state(target_nid, N_MEMORY)) {
-               while (!list_empty(folio_list)) {
-                       struct folio *folio = lru_to_folio(folio_list);
-
-                       list_del(&folio->lru);
-                       folio_putback_lru(folio);
-               }
-               return nr_migrated;
-       }
+                       !node_state(target_nid, N_MEMORY))
+               goto out;

        noreclaim_flag = memalloc_noreclaim_save();

@@ -424,6 +417,14 @@ unsigned long damon_migrate_pages(struct list_head *folio_list, int target_nid)

        memalloc_noreclaim_restore(noreclaim_flag);

+out:
+
+       while (!list_empty(folio_list)) {
+               struct folio *folio = lru_to_folio(folio_list);
+
+               list_del(&folio->lru);
+               folio_putback_lru(folio);
+       }
        return nr_migrated;
 }

This could be applied to the first patch.  And this patch can simply remove the
redundanty putback.

>  
> -- 
> 2.25.1

[1] https://lore.kernel.org/20260724060707.CE1F81F000E9@smtp.kernel.org


Thanks,
SJ

Re: [PATCH v2 1/2] mm/damon/ops-common: putback folios on invalid migrate nid

[SJ Park]

On Fri, 24 Jul 2026 14:01:34 +0800 dayou5941@163.com wrote:

> From: liyouhong <liyouhong@kylinos.cn>
> 
> damon_pa_migrate() and damos_va_migrate() isolate folios into a local list
> and then call damon_migrate_pages().  When target_nid is invalid (including
> the scheme default NUMA_NO_NODE / -1), damon_migrate_pages() returns early
> without putting the folios back to the LRU.
> 
> Callers then discard the list head while those folios remain isolated with
> an extra reference taken by folio_isolate_lru().  The pages stay off the
> LRU for as long as the mapping exists (anon active+inactive counts drop
> while RSS does not), and the leftover references can pin the pages after
> the mapping is gone.
> 
> Put the folios back on the invalid-nid path so ignored migration requests
> still return them to the LRU.
> 
> Fixes: 7e6c3130690a ("mm/damon/ops-common: ignore migration request to invalid nodes")
> Cc: <stable@vger.kernel.org>
> Assisted-by: Cursor:grok-4.5
> Signed-off-by: liyouhong <liyouhong@kylinos.cn>
> 
> ---
> v2:
> - Drop the putback helper for easier stable backport
> - Open-code putback only on the invalid-nid path
> - Keep the original indentation and Cc stable@
> - Add a separate cleanup as 2/2 (factor out damon_putback_folio_list)
> 
> ---
>  mm/damon/ops-common.c | 9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
> 
> diff --git a/mm/damon/ops-common.c b/mm/damon/ops-common.c
> index d1842e2b00ef..f5ded45fabd1 100644
> --- a/mm/damon/ops-common.c
> +++ b/mm/damon/ops-common.c
> @@ -394,8 +394,15 @@ unsigned long damon_migrate_pages(struct list_head *folio_list, int target_nid)
>  		return nr_migrated;
>  
>  	if (target_nid < 0 || target_nid >= MAX_NUMNODES ||
> -			!node_state(target_nid, N_MEMORY))
> +			!node_state(target_nid, N_MEMORY)) {
> +		while (!list_empty(folio_list)) {
> +			struct folio *folio = lru_to_folio(folio_list);
> +
> +			list_del(&folio->lru);
> +			folio_putback_lru(folio);
> +		}
>  		return nr_migrated;
> +	}

As I replied to the second patch, I'd suggest doing putback just before
returning.  Then this invalid target_nid handling can 'goto' there.


Thanks,
SJ

[...]

Re: [PATCH v2 2/2] mm/damon/ops-common: factor out damon_putback_folio_list()

[SJ Park]

On Fri, 24 Jul 2026 07:39:24 -0700 SJ Park <sj@kernel.org> wrote:

> 'get_maintainer.pl --nogit --nogit-fallback' suggests adding below recipients.
> I added them.  Please consier using get_maintainer.pl from the next time.
> 
> - linux-mm@kvack.org
> - linux-kernel@vger.kernel.org
> 
> On Fri, 24 Jul 2026 14:01:35 +0800 dayou5941@163.com wrote:
> 
> > From: liyouhong <liyouhong@kylinos.cn>
> > 
> > The putback loop is duplicated in damon_migrate_folio_list() and on the
> > invalid-nid path of damon_migrate_pages().  Factor it into a small helper
> > for readability.  No functional change.
> 
> This is not a hotfix.  I'd suggest sending this separately, not together with
> the first patch of this series.  Sending hotfix together with non-hotfix when
> they don't really need to be applied together only makes it complicated.
> 
> > 
> > Signed-off-by: liyouhong <liyouhong@kylinos.cn>
> > ---
> >  mm/damon/ops-common.c | 25 +++++++++++++------------
> >  1 file changed, 13 insertions(+), 12 deletions(-)
> > 
> > diff --git a/mm/damon/ops-common.c b/mm/damon/ops-common.c
> > index f5ded45fabd1..9a1e8aec5444 100644
> > --- a/mm/damon/ops-common.c
> > +++ b/mm/damon/ops-common.c
> > @@ -331,12 +331,22 @@ static unsigned int __damon_migrate_folio_list(
> >  	return nr_succeeded;
> >  }
> >  
> > +static void damon_putback_folio_list(struct list_head *folio_list)
> > +{
> > +	struct folio *folio;
> > +
> > +	while (!list_empty(folio_list)) {
> > +		folio = lru_to_folio(folio_list);
> > +		list_del(&folio->lru);
> > +		folio_putback_lru(folio);
> > +	}
> > +}
> > +
> >  static unsigned int damon_migrate_folio_list(struct list_head *folio_list,
> >  						struct pglist_data *pgdat,
> >  						int target_nid)
> >  {
> >  	unsigned int nr_migrated = 0;
> > -	struct folio *folio;

I forgot mentioning this breaks build, as Sashiko also pointed [1] out.  This
patch cannot be applied as-is.

[1] https://lore.kernel.org/20260724060707.CE1F81F000E9@smtp.kernel.org


Thanks,
SJ

> >  	LIST_HEAD(ret_folios);
> >  	LIST_HEAD(migrate_folios);
> >  
> > @@ -374,11 +384,7 @@ static unsigned int damon_migrate_folio_list(struct list_head *folio_list,
> >  
> >  	list_splice(&ret_folios, folio_list);
> >  
> > -	while (!list_empty(folio_list)) {
> > -		folio = lru_to_folio(folio_list);
> > -		list_del(&folio->lru);
> > -		folio_putback_lru(folio);
> > -	}
> > +	damon_putback_folio_list(folio_list);
> >  
> >  	return nr_migrated;
> >  }
> > @@ -395,12 +401,7 @@ unsigned long damon_migrate_pages(struct list_head *folio_list, int target_nid)
> >  
> >  	if (target_nid < 0 || target_nid >= MAX_NUMNODES ||
> >  			!node_state(target_nid, N_MEMORY)) {
> > -		while (!list_empty(folio_list)) {
> > -			struct folio *folio = lru_to_folio(folio_list);
> > -
> > -			list_del(&folio->lru);
> > -			folio_putback_lru(folio);
> > -		}
> > +		damon_putback_folio_list(folio_list);
> >  		return nr_migrated;
> >  	}
> 
> Looks better.  But, how about further simplifying it by moving the folios
> putback from damon_migrate_pages(), and doing that from damon_migrate_pages()?
> damon_migrate_pages() would do the putback always before returning, and the
> taregt_nid path will 'goto' the path.  E.g.,
> 
> --- a/mm/damon/ops-common.c
> +++ b/mm/damon/ops-common.c
> @@ -391,15 +391,8 @@ unsigned long damon_migrate_pages(struct list_head *folio_list, int target_nid)
>                 return nr_migrated;
> 
>         if (target_nid < 0 || target_nid >= MAX_NUMNODES ||
> -                       !node_state(target_nid, N_MEMORY)) {
> -               while (!list_empty(folio_list)) {
> -                       struct folio *folio = lru_to_folio(folio_list);
> -
> -                       list_del(&folio->lru);
> -                       folio_putback_lru(folio);
> -               }
> -               return nr_migrated;
> -       }
> +                       !node_state(target_nid, N_MEMORY))
> +               goto out;
> 
>         noreclaim_flag = memalloc_noreclaim_save();
> 
> @@ -424,6 +417,14 @@ unsigned long damon_migrate_pages(struct list_head *folio_list, int target_nid)
> 
>         memalloc_noreclaim_restore(noreclaim_flag);
> 
> +out:
> +
> +       while (!list_empty(folio_list)) {
> +               struct folio *folio = lru_to_folio(folio_list);
> +
> +               list_del(&folio->lru);
> +               folio_putback_lru(folio);
> +       }
>         return nr_migrated;
>  }
> 
> This could be applied to the first patch.  And this patch can simply remove the
> redundanty putback.
> 
> >  
> > -- 
> > 2.25.1
> 
> [1] https://lore.kernel.org/20260724060707.CE1F81F000E9@smtp.kernel.org
> 
> 
> Thanks,
> SJ
> 

Re: [PATCH v2 1/2] mm/damon/ops-common: putback folios on invalid migrate nid

[SJ Park]

On Fri, 24 Jul 2026 07:41:46 -0700 SJ Park <sj@kernel.org> wrote:

> On Fri, 24 Jul 2026 14:01:34 +0800 dayou5941@163.com wrote:
> 
> > From: liyouhong <liyouhong@kylinos.cn>
> > 
> > damon_pa_migrate() and damos_va_migrate() isolate folios into a local list
> > and then call damon_migrate_pages().  When target_nid is invalid (including
> > the scheme default NUMA_NO_NODE / -1), damon_migrate_pages() returns early
> > without putting the folios back to the LRU.
> > 
> > Callers then discard the list head while those folios remain isolated with
> > an extra reference taken by folio_isolate_lru().  The pages stay off the
> > LRU for as long as the mapping exists (anon active+inactive counts drop
> > while RSS does not), and the leftover references can pin the pages after
> > the mapping is gone.
> > 
> > Put the folios back on the invalid-nid path so ignored migration requests
> > still return them to the LRU.
> > 
> > Fixes: 7e6c3130690a ("mm/damon/ops-common: ignore migration request to invalid nodes")
> > Cc: <stable@vger.kernel.org>
> > Assisted-by: Cursor:grok-4.5
> > Signed-off-by: liyouhong <liyouhong@kylinos.cn>
> > 
> > ---
> > v2:
> > - Drop the putback helper for easier stable backport
> > - Open-code putback only on the invalid-nid path
> > - Keep the original indentation and Cc stable@
> > - Add a separate cleanup as 2/2 (factor out damon_putback_folio_list)
> > 
> > ---
> >  mm/damon/ops-common.c | 9 ++++++++-
> >  1 file changed, 8 insertions(+), 1 deletion(-)
> > 
> > diff --git a/mm/damon/ops-common.c b/mm/damon/ops-common.c
> > index d1842e2b00ef..f5ded45fabd1 100644
> > --- a/mm/damon/ops-common.c
> > +++ b/mm/damon/ops-common.c
> > @@ -394,8 +394,15 @@ unsigned long damon_migrate_pages(struct list_head *folio_list, int target_nid)
> >  		return nr_migrated;
> >  
> >  	if (target_nid < 0 || target_nid >= MAX_NUMNODES ||
> > -			!node_state(target_nid, N_MEMORY))
> > +			!node_state(target_nid, N_MEMORY)) {
> > +		while (!list_empty(folio_list)) {
> > +			struct folio *folio = lru_to_folio(folio_list);
> > +
> > +			list_del(&folio->lru);
> > +			folio_putback_lru(folio);
> > +		}
> >  		return nr_migrated;
> > +	}
> 
> As I replied to the second patch, I'd suggest doing putback just before
> returning.  Then this invalid target_nid handling can 'goto' there.

I now think that is a bad idea.  Sorry for confusing.  That will make readers
assume damon_migrate_folio_list() doesn't putback folios.  Please resend this
as-is, but without the second patch.


Thanks,
SJ

[...]