[PATCH v2 0/5] mm: memory_hotplug: improve do_migrate_range()

[PATCH v2 0/5] mm: memory_hotplug: improve do_migrate_range()

[Kefeng Wang]

Unify hwpoisoned page handling and isolation of HugeTLB/LRU/non-LRU
movable page, also convert to use folios in do_migrate_range().

v2:
- address comments from David(eg, fix HWPoison check/use a folio
  for pfn calculation firstly)
- fix lkp build errors for isolate_folio_to_list()
- drop unnecessary comments and don't grab one more ref for hugetlb 

Kefeng Wang (5):
  mm: memory_hotplug: remove head variable in do_migrate_range()
  mm: memory-failure: add unmap_posioned_folio()
  mm: memory_hotplug: check hwpoisoned page firstly in
    do_migrate_range()
  mm: migrate: add isolate_folio_to_list()
  mm: memory_hotplug: unify Huge/LRU/non-LRU movable folio isolation

 include/linux/migrate.h |  3 ++
 mm/internal.h           |  9 +++++
 mm/memory-failure.c     | 89 +++++++++++++++++------------------------
 mm/memory_hotplug.c     | 62 ++++++++++++++--------------
 mm/migrate.c            | 27 +++++++++++++
 5 files changed, 105 insertions(+), 85 deletions(-)

-- 
2.27.0

[PATCH v2 1/5] mm: memory_hotplug: remove head variable in do_migrate_range()

[Kefeng Wang]

Directly use a folio for HugeTLB and THP when calculate the next pfn,
then remove unused head variable.

Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
---
 mm/memory_hotplug.c | 22 ++++++++++++++--------
 1 file changed, 14 insertions(+), 8 deletions(-)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index df291f2e509d..7c674cec3c18 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -1773,7 +1773,7 @@ static int scan_movable_pages(unsigned long start, unsigned long end,
 static void do_migrate_range(unsigned long start_pfn, unsigned long end_pfn)
 {
 	unsigned long pfn;
-	struct page *page, *head;
+	struct page *page;
 	LIST_HEAD(source);
 	static DEFINE_RATELIMIT_STATE(migrate_rs, DEFAULT_RATELIMIT_INTERVAL,
 				      DEFAULT_RATELIMIT_BURST);
@@ -1786,14 +1786,20 @@ static void do_migrate_range(unsigned long start_pfn, unsigned long end_pfn)
 			continue;
 		page = pfn_to_page(pfn);
 		folio = page_folio(page);
-		head = &folio->page;
 
-		if (PageHuge(page)) {
-			pfn = page_to_pfn(head) + compound_nr(head) - 1;
-			isolate_hugetlb(folio, &source);
-			continue;
-		} else if (PageTransHuge(page))
-			pfn = page_to_pfn(head) + thp_nr_pages(page) - 1;
+		/*
+		 * The folio hold no reference or lock, it might be changed
+		 * concurrently(eg, split), folio_nr_pages() may read garbage,
+		 * but out loop could handle that as it revisits the split
+		 * folio later.
+		 */
+		if (folio_test_large(folio)) {
+			pfn = folio_pfn(folio) + folio_nr_pages(folio) - 1;
+			if (folio_test_hugetlb(folio)) {
+				isolate_hugetlb(folio);
+				continue;
+			}
+		}
 
 		/*
 		 * HWPoison pages have elevated reference counts so the migration would
-- 
2.27.0

[PATCH v2 2/5] mm: memory-failure: add unmap_posioned_folio()

[Kefeng Wang]

Add unmap_posioned_folio() helper which will be reused by
do_migrate_range() from memory hotplug soon.

Acked-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
---
 mm/internal.h       |  9 +++++++++
 mm/memory-failure.c | 43 ++++++++++++++++++++++++++-----------------
 2 files changed, 35 insertions(+), 17 deletions(-)

diff --git a/mm/internal.h b/mm/internal.h
index adbf8c88c9df..5b80c65f82b6 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -1054,6 +1054,8 @@ static inline int find_next_best_node(int node, nodemask_t *used_node_mask)
 /*
  * mm/memory-failure.c
  */
+#ifdef CONFIG_MEMORY_FAILURE
+int unmap_posioned_folio(struct folio *folio, enum ttu_flags ttu);
 void shake_folio(struct folio *folio);
 extern int hwpoison_filter(struct page *p);
 
@@ -1074,6 +1076,13 @@ void add_to_kill_ksm(struct task_struct *tsk, struct page *p,
 		     unsigned long ksm_addr);
 unsigned long page_mapped_in_vma(struct page *page, struct vm_area_struct *vma);
 
+#else
+static inline int unmap_posioned_folio(struct folio *folio, enum ttu_flags ttu)
+{
+	return 0;
+}
+#endif
+
 extern unsigned long  __must_check vm_mmap_pgoff(struct file *, unsigned long,
         unsigned long, unsigned long,
         unsigned long, unsigned long);
diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 353254537b54..93848330de1f 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -1554,6 +1554,30 @@ static int get_hwpoison_page(struct page *p, unsigned long flags)
 	return ret;
 }
 
+int unmap_posioned_folio(struct folio *folio, enum ttu_flags ttu)
+{
+	if (folio_test_hugetlb(folio) && !folio_test_anon(folio)) {
+		struct address_space *mapping;
+		/*
+		 * For hugetlb pages in shared mappings, try_to_unmap
+		 * could potentially call huge_pmd_unshare.  Because of
+		 * this, take semaphore in write mode here and set
+		 * TTU_RMAP_LOCKED to indicate we have taken the lock
+		 * at this higher level.
+		 */
+		mapping = hugetlb_folio_mapping_lock_write(folio);
+		if (!mapping)
+			return -EAGAIN;
+
+		try_to_unmap(folio, ttu|TTU_RMAP_LOCKED);
+		i_mmap_unlock_write(mapping);
+	} else {
+		try_to_unmap(folio, ttu);
+	}
+
+	return 0;
+}
+
 /*
  * Do all that is necessary to remove user space mappings. Unmap
  * the pages and send SIGBUS to the processes if the data was dirty.
@@ -1615,23 +1639,8 @@ static bool hwpoison_user_mappings(struct folio *folio, struct page *p,
 	 */
 	collect_procs(folio, p, &tokill, flags & MF_ACTION_REQUIRED);
 
-	if (folio_test_hugetlb(folio) && !folio_test_anon(folio)) {
-		/*
-		 * For hugetlb pages in shared mappings, try_to_unmap
-		 * could potentially call huge_pmd_unshare.  Because of
-		 * this, take semaphore in write mode here and set
-		 * TTU_RMAP_LOCKED to indicate we have taken the lock
-		 * at this higher level.
-		 */
-		mapping = hugetlb_folio_mapping_lock_write(folio);
-		if (mapping) {
-			try_to_unmap(folio, ttu|TTU_RMAP_LOCKED);
-			i_mmap_unlock_write(mapping);
-		} else
-			pr_info("%#lx: could not lock mapping for mapped huge page\n", pfn);
-	} else {
-		try_to_unmap(folio, ttu);
-	}
+	if (unmap_posioned_folio(folio, ttu))
+		pr_info("%#lx: could not lock mapping for mapped huge page\n", pfn);
 
 	unmap_success = !folio_mapped(folio);
 	if (!unmap_success)
-- 
2.27.0

[PATCH v2 3/5] mm: memory_hotplug: check hwpoisoned page firstly in do_migrate_range()

[Kefeng Wang]

The commit b15c87263a69 ("hwpoison, memory_hotplug: allow hwpoisoned
pages to be offlined") don't handle the hugetlb pages, the endless
loop still occur if offline a hwpoison hugetlb, luckly, with the
commit e591ef7d96d6 ("mm,hwpoison,hugetlb,memory_hotplug: hotremove
memory section with hwpoisoned hugepage") section with hwpoisoned
hugepage"), the HPageMigratable of hugetlb page will be clear, and
the hwpoison hugetlb page will be skipped in scan_movable_pages(),
so the endless loop issue is fixed.

However if the HPageMigratable() check passed(without reference and
lock), the hugetlb page may be hwpoisoned, it won't cause issue since
the hwpoisoned page will be handled correctly in the next movable
pages scan loop, and it will be isolated in do_migrate_range() but
fails to migrate. In order to avoid the unnecessary isolation and
unify all hwpoisoned page handling, let's unconditionally check hwpoison
firstly, and if it is a hwpoisoned hugetlb page, try to unmap it as
the catch all safety net like normal page does.

Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
---
 mm/memory_hotplug.c | 17 +++++++++--------
 1 file changed, 9 insertions(+), 8 deletions(-)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 7c674cec3c18..02a0d4fbc3fe 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -1793,13 +1793,8 @@ static void do_migrate_range(unsigned long start_pfn, unsigned long end_pfn)
 		 * but out loop could handle that as it revisits the split
 		 * folio later.
 		 */
-		if (folio_test_large(folio)) {
+		if (folio_test_large(folio))
 			pfn = folio_pfn(folio) + folio_nr_pages(folio) - 1;
-			if (folio_test_hugetlb(folio)) {
-				isolate_hugetlb(folio);
-				continue;
-			}
-		}
 
 		/*
 		 * HWPoison pages have elevated reference counts so the migration would
@@ -1808,11 +1803,17 @@ static void do_migrate_range(unsigned long start_pfn, unsigned long end_pfn)
 		 * (e.g. current hwpoison implementation doesn't unmap KSM pages but keep
 		 * the unmap as the catch all safety net).
 		 */
-		if (PageHWPoison(page)) {
+		if (folio_test_hwpoison(folio) ||
+		    (folio_test_large(folio) && folio_test_has_hwpoisoned(folio))) {
 			if (WARN_ON(folio_test_lru(folio)))
 				folio_isolate_lru(folio);
 			if (folio_mapped(folio))
-				try_to_unmap(folio, TTU_IGNORE_MLOCK);
+				unmap_posioned_folio(folio, TTU_IGNORE_MLOCK);
+			continue;
+		}
+
+		if (folio_test_hugetlb(folio)) {
+			isolate_hugetlb(folio, &source);
 			continue;
 		}
 
-- 
2.27.0

[PATCH v2 4/5] mm: migrate: add isolate_folio_to_list()

[Kefeng Wang]

Add isolate_folio_to_list() helper to try to isolate HugeTLB,
no-LRU movable and LRU folios to a list, which will be reused by
do_migrate_range() from memory hotplug soon, also drop the
mf_isolate_folio() since we could directly use new helper in
the soft_offline_in_use_page().

Acked-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
---
 include/linux/migrate.h |  3 +++
 mm/memory-failure.c     | 46 ++++++++++-------------------------------
 mm/migrate.c            | 27 ++++++++++++++++++++++++
 3 files changed, 41 insertions(+), 35 deletions(-)

diff --git a/include/linux/migrate.h b/include/linux/migrate.h
index 644be30b69c8..002e49b2ebd9 100644
--- a/include/linux/migrate.h
+++ b/include/linux/migrate.h
@@ -70,6 +70,7 @@ int migrate_pages(struct list_head *l, new_folio_t new, free_folio_t free,
 		  unsigned int *ret_succeeded);
 struct folio *alloc_migration_target(struct folio *src, unsigned long private);
 bool isolate_movable_page(struct page *page, isolate_mode_t mode);
+bool isolate_folio_to_list(struct folio *folio, struct list_head *list);
 
 int migrate_huge_page_move_mapping(struct address_space *mapping,
 		struct folio *dst, struct folio *src);
@@ -91,6 +92,8 @@ static inline struct folio *alloc_migration_target(struct folio *src,
 	{ return NULL; }
 static inline bool isolate_movable_page(struct page *page, isolate_mode_t mode)
 	{ return false; }
+static inline bool isolate_folio_to_list(struct folio *folio, struct list_head *list)
+	{ return false; }
 
 static inline int migrate_huge_page_move_mapping(struct address_space *mapping,
 				  struct folio *dst, struct folio *src)
diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 93848330de1f..d8298017bd99 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -2659,40 +2659,6 @@ EXPORT_SYMBOL(unpoison_memory);
 #undef pr_fmt
 #define pr_fmt(fmt) "Soft offline: " fmt
 
-static bool mf_isolate_folio(struct folio *folio, struct list_head *pagelist)
-{
-	bool isolated = false;
-
-	if (folio_test_hugetlb(folio)) {
-		isolated = isolate_hugetlb(folio, pagelist);
-	} else {
-		bool lru = !__folio_test_movable(folio);
-
-		if (lru)
-			isolated = folio_isolate_lru(folio);
-		else
-			isolated = isolate_movable_page(&folio->page,
-							ISOLATE_UNEVICTABLE);
-
-		if (isolated) {
-			list_add(&folio->lru, pagelist);
-			if (lru)
-				node_stat_add_folio(folio, NR_ISOLATED_ANON +
-						    folio_is_file_lru(folio));
-		}
-	}
-
-	/*
-	 * If we succeed to isolate the folio, we grabbed another refcount on
-	 * the folio, so we can safely drop the one we got from get_any_page().
-	 * If we failed to isolate the folio, it means that we cannot go further
-	 * and we will return an error, so drop the reference we got from
-	 * get_any_page() as well.
-	 */
-	folio_put(folio);
-	return isolated;
-}
-
 /*
  * soft_offline_in_use_page handles hugetlb-pages and non-hugetlb pages.
  * If the page is a non-dirty unmapped page-cache page, it simply invalidates.
@@ -2744,7 +2710,7 @@ static int soft_offline_in_use_page(struct page *page)
 		return 0;
 	}
 
-	if (mf_isolate_folio(folio, &pagelist)) {
+	if (isolate_folio_to_list(folio, &pagelist)) {
 		ret = migrate_pages(&pagelist, alloc_migration_target, NULL,
 			(unsigned long)&mtc, MIGRATE_SYNC, MR_MEMORY_FAILURE, NULL);
 		if (!ret) {
@@ -2766,6 +2732,16 @@ static int soft_offline_in_use_page(struct page *page)
 			pfn, msg_page[huge], page_count(page), &page->flags);
 		ret = -EBUSY;
 	}
+
+	/*
+	 * If we succeed to isolate the folio, we grabbed another refcount on
+	 * the folio, so we can safely drop the one we got from get_any_page().
+	 * If we failed to isolate the folio, it means that we cannot go further
+	 * and we will return an error, so drop the reference we got from
+	 * get_any_page() as well.
+	 */
+	folio_put(folio);
+
 	return ret;
 }
 
diff --git a/mm/migrate.c b/mm/migrate.c
index 6e32098ac2dc..7b7b5b16e610 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -178,6 +178,33 @@ void putback_movable_pages(struct list_head *l)
 	}
 }
 
+/* Must be called with an elevated refcount on the non-hugetlb folio */
+bool isolate_folio_to_list(struct folio *folio, struct list_head *list)
+{
+	bool isolated = false;
+
+	if (folio_test_hugetlb(folio)) {
+		isolated = isolate_hugetlb(folio, list);
+	} else {
+		bool lru = !__folio_test_movable(folio);
+
+		if (lru)
+			isolated = folio_isolate_lru(folio);
+		else
+			isolated = isolate_movable_page(&folio->page,
+							ISOLATE_UNEVICTABLE);
+
+		if (isolated) {
+			list_add(&folio->lru, list);
+			if (lru)
+				node_stat_add_folio(folio, NR_ISOLATED_ANON +
+						    folio_is_file_lru(folio));
+		}
+	}
+
+	return isolated;
+}
+
 static bool try_to_map_unused_to_zeropage(struct page_vma_mapped_walk *pvmw,
 					  struct folio *folio,
 					  unsigned long idx)
-- 
2.27.0

[PATCH v2 5/5] mm: memory_hotplug: unify Huge/LRU/non-LRU movable folio isolation

[Kefeng Wang]

Use the isolate_folio_to_list() to unify hugetlb/LRU/non-LRU
folio isolation, which cleanup code a bit and save a few calls
to compound_head().

Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
---
 mm/memory_hotplug.c | 45 +++++++++++++++++----------------------------
 1 file changed, 17 insertions(+), 28 deletions(-)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 02a0d4fbc3fe..cc9c16db2f8c 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -1773,14 +1773,14 @@ static int scan_movable_pages(unsigned long start, unsigned long end,
 static void do_migrate_range(unsigned long start_pfn, unsigned long end_pfn)
 {
 	unsigned long pfn;
-	struct page *page;
 	LIST_HEAD(source);
+	struct folio *folio;
 	static DEFINE_RATELIMIT_STATE(migrate_rs, DEFAULT_RATELIMIT_INTERVAL,
 				      DEFAULT_RATELIMIT_BURST);
 
 	for (pfn = start_pfn; pfn < end_pfn; pfn++) {
-		struct folio *folio;
-		bool isolated;
+		struct page *page;
+		bool huge;
 
 		if (!pfn_valid(pfn))
 			continue;
@@ -1812,34 +1812,22 @@ static void do_migrate_range(unsigned long start_pfn, unsigned long end_pfn)
 			continue;
 		}
 
-		if (folio_test_hugetlb(folio)) {
-			isolate_hugetlb(folio, &source);
-			continue;
+		huge = folio_test_hugetlb(folio);
+		if (!huge) {
+			folio = folio_get_nontail_page(page);
+			if (!folio)
+				continue;
 		}
 
-		if (!get_page_unless_zero(page))
-			continue;
-		/*
-		 * We can skip free pages. And we can deal with pages on
-		 * LRU and non-lru movable pages.
-		 */
-		if (PageLRU(page))
-			isolated = isolate_lru_page(page);
-		else
-			isolated = isolate_movable_page(page, ISOLATE_UNEVICTABLE);
-		if (isolated) {
-			list_add_tail(&page->lru, &source);
-			if (!__PageMovable(page))
-				inc_node_page_state(page, NR_ISOLATED_ANON +
-						    page_is_file_lru(page));
-
-		} else {
+		if (!isolate_folio_to_list(folio, &source)) {
 			if (__ratelimit(&migrate_rs)) {
 				pr_warn("failed to isolate pfn %lx\n", pfn);
 				dump_page(page, "isolation failed");
 			}
 		}
-		put_page(page);
+
+		if (!huge)
+			folio_put(folio);
 	}
 	if (!list_empty(&source)) {
 		nodemask_t nmask = node_states[N_MEMORY];
@@ -1854,7 +1842,7 @@ static void do_migrate_range(unsigned long start_pfn, unsigned long end_pfn)
 		 * We have checked that migration range is on a single zone so
 		 * we can use the nid of the first page to all the others.
 		 */
-		mtc.nid = page_to_nid(list_first_entry(&source, struct page, lru));
+		mtc.nid = folio_nid(list_first_entry(&source, struct folio, lru));
 
 		/*
 		 * try to allocate from a different node but reuse this node
@@ -1867,11 +1855,12 @@ static void do_migrate_range(unsigned long start_pfn, unsigned long end_pfn)
 		ret = migrate_pages(&source, alloc_migration_target, NULL,
 			(unsigned long)&mtc, MIGRATE_SYNC, MR_MEMORY_HOTPLUG, NULL);
 		if (ret) {
-			list_for_each_entry(page, &source, lru) {
+			list_for_each_entry(folio, &source, lru) {
 				if (__ratelimit(&migrate_rs)) {
 					pr_warn("migrating pfn %lx failed ret:%d\n",
-						page_to_pfn(page), ret);
-					dump_page(page, "migration failure");
+						folio_pfn(folio), ret);
+					dump_page(&folio->page,
+						  "migration failure");
 				}
 			}
 			putback_movable_pages(&source);
-- 
2.27.0

Re: [PATCH v2 0/5] mm: memory_hotplug: improve do_migrate_range()

[Kefeng Wang]

Sorry for the noise, please ignore this one, will resend.

On 2024/8/16 17:04, Kefeng Wang wrote:
> Unify hwpoisoned page handling and isolation of HugeTLB/LRU/non-LRU
> movable page, also convert to use folios in do_migrate_range().
> 
> v2:
> - address comments from David(eg, fix HWPoison check/use a folio
>    for pfn calculation firstly)
> - fix lkp build errors for isolate_folio_to_list()
> - drop unnecessary comments and don't grab one more ref for hugetlb
> 
> Kefeng Wang (5):
>    mm: memory_hotplug: remove head variable in do_migrate_range()
>    mm: memory-failure: add unmap_posioned_folio()
>    mm: memory_hotplug: check hwpoisoned page firstly in
>      do_migrate_range()
>    mm: migrate: add isolate_folio_to_list()
>    mm: memory_hotplug: unify Huge/LRU/non-LRU movable folio isolation
> 
>   include/linux/migrate.h |  3 ++
>   mm/internal.h           |  9 +++++
>   mm/memory-failure.c     | 89 +++++++++++++++++------------------------
>   mm/memory_hotplug.c     | 62 ++++++++++++++--------------
>   mm/migrate.c            | 27 +++++++++++++
>   5 files changed, 105 insertions(+), 85 deletions(-)
>