[PATCH 0/3] Some cleanups for page isolation

[PATCH 0/3] Some cleanups for page isolation

[Baolin Wang]

Hi,

The page isolation functions did not return a boolean to indicate
success or not, instead it will return a negative error when failed
to isolate a page. So it's better to check the negative error explicitly
for isolation to make the code more clear per Linus's suggestion in [1].

No functional changes intended in this patch series.

[1] https://lore.kernel.org/all/CAHk-=wiBrY+O-4=2mrbVyxR+hOqfdJ=Do6xoucfJ9_5az01L4Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org/

Baolin Wang (3):
  mm: check negative error of folio_isolate_lru() when failed to isolate
    a folio
  mm: check negative error of isolate_lru_page() when failed to isolate
    a page
  mm: mempolicy: check negative error of isolate_hugetlb() when failed
    to isolate a hugetlb

 mm/damon/paddr.c    | 2 +-
 mm/gup.c            | 2 +-
 mm/khugepaged.c     | 4 ++--
 mm/memcontrol.c     | 2 +-
 mm/mempolicy.c      | 2 +-
 mm/migrate.c        | 4 ++--
 mm/migrate_device.c | 2 +-
 7 files changed, 9 insertions(+), 9 deletions(-)

-- 
2.27.0

[PATCH 1/3] mm: check negative error of folio_isolate_lru() when failed to isolate a folio

[Baolin Wang]

The folio_isolate_lru() will return a negative error if failed to isolate
a folio from its LRU list, thus better to check the negative error to
make the code more clear per Linus's suggestion[1].

No functional changes.

[1] https://lore.kernel.org/all/CAHk-=wiBrY+O-4=2mrbVyxR+hOqfdJ=Do6xoucfJ9_5az01L4Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org/

Signed-off-by: Baolin Wang <baolin.wang-KPsoFbNs7GizrGE5bRqYAgC/G2K4zDHf@public.gmane.org>
---
 mm/damon/paddr.c | 2 +-
 mm/gup.c         | 2 +-
 mm/khugepaged.c  | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/mm/damon/paddr.c b/mm/damon/paddr.c
index b4df9b9bcc0a..ebf3dd084af2 100644
--- a/mm/damon/paddr.c
+++ b/mm/damon/paddr.c
@@ -246,7 +246,7 @@ static unsigned long damon_pa_pageout(struct damon_region *r, struct damos *s)
 
 		folio_clear_referenced(folio);
 		folio_test_clear_young(folio);
-		if (folio_isolate_lru(folio)) {
+		if (folio_isolate_lru(folio) < 0) {
 			folio_put(folio);
 			continue;
 		}
diff --git a/mm/gup.c b/mm/gup.c
index b0885f70579c..7fdddd43663c 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1939,7 +1939,7 @@ static unsigned long collect_longterm_unpinnable_pages(
 			drain_allow = false;
 		}
 
-		if (folio_isolate_lru(folio))
+		if (folio_isolate_lru(folio) < 0)
 			continue;
 
 		list_add_tail(&folio->lru, movable_page_list);
diff --git a/mm/khugepaged.c b/mm/khugepaged.c
index a5d32231bfad..463dfca4b841 100644
--- a/mm/khugepaged.c
+++ b/mm/khugepaged.c
@@ -2047,7 +2047,7 @@ static int collapse_file(struct mm_struct *mm, unsigned long addr,
 			goto out_unlock;
 		}
 
-		if (folio_isolate_lru(folio)) {
+		if (folio_isolate_lru(folio) < 0) {
 			result = SCAN_DEL_PAGE_LRU;
 			goto out_unlock;
 		}
-- 
2.27.0

[PATCH 2/3] mm: check negative error of isolate_lru_page() when failed to isolate a page

[Baolin Wang]

Better to check the negative error of isolate_lru_page() when failed to
isolate a page from its LRU, which makes the code more clear.

No functional changes.

Signed-off-by: Baolin Wang <baolin.wang-KPsoFbNs7GizrGE5bRqYAgC/G2K4zDHf@public.gmane.org>
---
 mm/khugepaged.c     | 2 +-
 mm/memcontrol.c     | 2 +-
 mm/migrate.c        | 4 ++--
 mm/migrate_device.c | 2 +-
 4 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/mm/khugepaged.c b/mm/khugepaged.c
index 463dfca4b841..99a2d993b51c 100644
--- a/mm/khugepaged.c
+++ b/mm/khugepaged.c
@@ -659,7 +659,7 @@ static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
 		 * Isolate the page to avoid collapsing an hugepage
 		 * currently in use by the VM.
 		 */
-		if (isolate_lru_page(page)) {
+		if (isolate_lru_page(page) < 0) {
 			unlock_page(page);
 			result = SCAN_DEL_PAGE_LRU;
 			goto out;
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 17335459d8dc..33cad66a4132 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -6226,7 +6226,7 @@ static int mem_cgroup_move_charge_pte_range(pmd_t *pmd,
 			 */
 			if (PageTransCompound(page))
 				goto put;
-			if (!device && isolate_lru_page(page))
+			if (!device && isolate_lru_page(page) < 0)
 				goto put;
 			if (!mem_cgroup_move_account(page, false,
 						mc.from, mc.to)) {
diff --git a/mm/migrate.c b/mm/migrate.c
index ef68a1aff35c..eaae39564f86 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -2135,7 +2135,7 @@ static int add_page_for_migration(struct mm_struct *mm, unsigned long addr,
 
 		head = compound_head(page);
 		err = isolate_lru_page(head);
-		if (err)
+		if (err < 0)
 			goto out_putpage;
 
 		err = 1;
@@ -2541,7 +2541,7 @@ static int numamigrate_isolate_page(pg_data_t *pgdat, struct page *page)
 		return 0;
 	}
 
-	if (isolate_lru_page(page))
+	if (isolate_lru_page(page) < 0)
 		return 0;
 
 	mod_node_page_state(page_pgdat(page), NR_ISOLATED_ANON + page_is_file_lru(page),
diff --git a/mm/migrate_device.c b/mm/migrate_device.c
index 6c3740318a98..1ab56b5b7914 100644
--- a/mm/migrate_device.c
+++ b/mm/migrate_device.c
@@ -388,7 +388,7 @@ static unsigned long migrate_device_unmap(unsigned long *src_pfns,
 				allow_drain = false;
 			}
 
-			if (isolate_lru_page(page)) {
+			if (isolate_lru_page(page) < 0) {
 				src_pfns[i] &= ~MIGRATE_PFN_MIGRATE;
 				restore++;
 				continue;
-- 
2.27.0

[PATCH 3/3] mm: mempolicy: check negative error of isolate_hugetlb() when failed to isolate a hugetlb

[Baolin Wang]

Better to check the negative error of isolate_hugetlb() when failed to
isolate a hugetlb page, which makes the code more clear.

No functional changes.

Signed-off-by: Baolin Wang <baolin.wang-KPsoFbNs7GizrGE5bRqYAgC/G2K4zDHf@public.gmane.org>
---
 mm/mempolicy.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/mm/mempolicy.c b/mm/mempolicy.c
index 0919c7a719d4..e3d87b21516b 100644
--- a/mm/mempolicy.c
+++ b/mm/mempolicy.c
@@ -609,7 +609,7 @@ static int queue_folios_hugetlb(pte_t *pte, unsigned long hmask,
 	if (flags & (MPOL_MF_MOVE_ALL) ||
 	    (flags & MPOL_MF_MOVE && folio_estimated_sharers(folio) == 1 &&
 	     !hugetlb_pmd_shared(pte))) {
-		if (isolate_hugetlb(folio, qp->pagelist) &&
+		if (isolate_hugetlb(folio, qp->pagelist)  < 0 &&
 			(flags & MPOL_MF_STRICT))
 			/*
 			 * Failed to isolate folio but allow migrating pages
-- 
2.27.0

Re: [PATCH 0/3] Some cleanups for page isolation

[Matthew Wilcox]

On Tue, Feb 14, 2023 at 11:18:05AM +0800, Baolin Wang wrote:
> The page isolation functions did not return a boolean to indicate
> success or not, instead it will return a negative error when failed
> to isolate a page. So it's better to check the negative error explicitly
> for isolation to make the code more clear per Linus's suggestion in [1].

Only one caller of isolate_lru_page() or folio_isolate_lru() actually
uses the errno.  And the errno can only be 0 or -EBUSY.  It'd be
better to change the three functions to return bool and fix
add_page_for_migration() to set the errno to -EBUSY itself.

Re: [PATCH 0/3] Some cleanups for page isolation

[Baolin Wang]

On 2/14/2023 12:50 PM, Matthew Wilcox wrote:
> On Tue, Feb 14, 2023 at 11:18:05AM +0800, Baolin Wang wrote:
>> The page isolation functions did not return a boolean to indicate
>> success or not, instead it will return a negative error when failed
>> to isolate a page. So it's better to check the negative error explicitly
>> for isolation to make the code more clear per Linus's suggestion in [1].
> 
> Only one caller of isolate_lru_page() or folio_isolate_lru() actually
> uses the errno.  And the errno can only be 0 or -EBUSY.  It'd be
> better to change the three functions to return bool and fix
> add_page_for_migration() to set the errno to -EBUSY itself.

Sounds reasonable to me, and I can change them to return bool in next 
version. Thanks.