[PATCH v3 00/11] Some cleanups for memory-failure

[PATCH v3 00/11] Some cleanups for memory-failure

[Matthew Wilcox (Oracle)]

A lot of folio conversions, plus some other simplifications.

v3:
 - Rebase on next-20240412
 - Reinstate missing hunks of "Return the address from page_mapped_in_vma()"
 - Accumulate R-b tags

Matthew Wilcox (Oracle) (11):
  mm/memory-failure: Remove fsdax_pgoff argument from __add_to_kill
  mm/memory-failure: Pass addr to __add_to_kill()
  mm: Return the address from page_mapped_in_vma()
  mm: Make page_mapped_in_vma conditional on CONFIG_MEMORY_FAILURE
  mm/memory-failure: Convert shake_page() to shake_folio()
  mm: Convert hugetlb_page_mapping_lock_write to folio
  mm/memory-failure: Convert memory_failure() to use a folio
  mm/memory-failure: Convert hwpoison_user_mappings to take a folio
  mm/memory-failure: Add some folio conversions to unpoison_memory
  mm/memory-failure: Use folio functions throughout collect_procs()
  mm/memory-failure: Pass the folio to collect_procs_ksm()

 include/linux/hugetlb.h |   6 +-
 include/linux/ksm.h     |  14 +---
 include/linux/mm.h      |   1 -
 include/linux/rmap.h    |   2 +-
 mm/hugetlb.c            |   6 +-
 mm/hwpoison-inject.c    |  11 +--
 mm/internal.h           |   1 +
 mm/ksm.c                |   5 +-
 mm/memory-failure.c     | 150 +++++++++++++++++++++-------------------
 mm/migrate.c            |   2 +-
 mm/page_vma_mapped.c    |  18 +++--
 11 files changed, 108 insertions(+), 108 deletions(-)

-- 
2.43.0

[PATCH v3 01/11] mm/memory-failure: Remove fsdax_pgoff argument from __add_to_kill

[Matthew Wilcox (Oracle)]

Unify the KSM and DAX codepaths by calculating the addr in
add_to_kill_fsdax() instead of telling __add_to_kill() to calculate it.

Acked-by: Miaohe Lin <linmiaohe@huawei.com>
Reviewed-by: Jane Chu <jane.chu@oracle.com>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/memory-failure.c | 27 +++++++++------------------
 1 file changed, 9 insertions(+), 18 deletions(-)

diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index ee2f4b8905ef..8adf233837bf 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -421,21 +421,13 @@ static unsigned long dev_pagemap_mapping_shift(struct vm_area_struct *vma,
  * not much we can do.	We just print a message and ignore otherwise.
  */
 
-#define FSDAX_INVALID_PGOFF ULONG_MAX
-
 /*
  * Schedule a process for later kill.
  * Uses GFP_ATOMIC allocations to avoid potential recursions in the VM.
- *
- * Note: @fsdax_pgoff is used only when @p is a fsdax page and a
- * filesystem with a memory failure handler has claimed the
- * memory_failure event. In all other cases, page->index and
- * page->mapping are sufficient for mapping the page back to its
- * corresponding user virtual address.
  */
 static void __add_to_kill(struct task_struct *tsk, struct page *p,
 			  struct vm_area_struct *vma, struct list_head *to_kill,
-			  unsigned long ksm_addr, pgoff_t fsdax_pgoff)
+			  unsigned long addr)
 {
 	struct to_kill *tk;
 
@@ -445,12 +437,10 @@ static void __add_to_kill(struct task_struct *tsk, struct page *p,
 		return;
 	}
 
-	tk->addr = ksm_addr ? ksm_addr : page_address_in_vma(p, vma);
-	if (is_zone_device_page(p)) {
-		if (fsdax_pgoff != FSDAX_INVALID_PGOFF)
-			tk->addr = vma_address(vma, fsdax_pgoff, 1);
+	tk->addr = addr ? addr : page_address_in_vma(p, vma);
+	if (is_zone_device_page(p))
 		tk->size_shift = dev_pagemap_mapping_shift(vma, tk->addr);
-	} else
+	else
 		tk->size_shift = page_shift(compound_head(p));
 
 	/*
@@ -480,7 +470,7 @@ static void add_to_kill_anon_file(struct task_struct *tsk, struct page *p,
 				  struct vm_area_struct *vma,
 				  struct list_head *to_kill)
 {
-	__add_to_kill(tsk, p, vma, to_kill, 0, FSDAX_INVALID_PGOFF);
+	__add_to_kill(tsk, p, vma, to_kill, 0);
 }
 
 #ifdef CONFIG_KSM
@@ -498,10 +488,10 @@ static bool task_in_to_kill_list(struct list_head *to_kill,
 }
 void add_to_kill_ksm(struct task_struct *tsk, struct page *p,
 		     struct vm_area_struct *vma, struct list_head *to_kill,
-		     unsigned long ksm_addr)
+		     unsigned long addr)
 {
 	if (!task_in_to_kill_list(to_kill, tsk))
-		__add_to_kill(tsk, p, vma, to_kill, ksm_addr, FSDAX_INVALID_PGOFF);
+		__add_to_kill(tsk, p, vma, to_kill, addr);
 }
 #endif
 /*
@@ -675,7 +665,8 @@ static void add_to_kill_fsdax(struct task_struct *tsk, struct page *p,
 			      struct vm_area_struct *vma,
 			      struct list_head *to_kill, pgoff_t pgoff)
 {
-	__add_to_kill(tsk, p, vma, to_kill, 0, pgoff);
+	unsigned long addr = vma_address(vma, pgoff, 1);
+	__add_to_kill(tsk, p, vma, to_kill, addr);
 }
 
 /*
-- 
2.43.0

[PATCH v3 02/11] mm/memory-failure: Pass addr to __add_to_kill()

[Matthew Wilcox (Oracle)]

Handle anon/file folios the same way as KSM & DAX folios by passing in
the address.

Acked-by: Miaohe Lin <linmiaohe@huawei.com>
Reviewed-by: Jane Chu <jane.chu@oracle.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/memory-failure.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 8adf233837bf..aec407788df1 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -437,7 +437,7 @@ static void __add_to_kill(struct task_struct *tsk, struct page *p,
 		return;
 	}
 
-	tk->addr = addr ? addr : page_address_in_vma(p, vma);
+	tk->addr = addr;
 	if (is_zone_device_page(p))
 		tk->size_shift = dev_pagemap_mapping_shift(vma, tk->addr);
 	else
@@ -470,7 +470,8 @@ static void add_to_kill_anon_file(struct task_struct *tsk, struct page *p,
 				  struct vm_area_struct *vma,
 				  struct list_head *to_kill)
 {
-	__add_to_kill(tsk, p, vma, to_kill, 0);
+	unsigned long addr = page_address_in_vma(p, vma);
+	__add_to_kill(tsk, p, vma, to_kill, addr);
 }
 
 #ifdef CONFIG_KSM
@@ -486,6 +487,7 @@ static bool task_in_to_kill_list(struct list_head *to_kill,
 
 	return false;
 }
+
 void add_to_kill_ksm(struct task_struct *tsk, struct page *p,
 		     struct vm_area_struct *vma, struct list_head *to_kill,
 		     unsigned long addr)
-- 
2.43.0

[PATCH v3 03/11] mm: Return the address from page_mapped_in_vma()

[Matthew Wilcox (Oracle)]

The only user of this function calls page_address_in_vma() immediately
after page_mapped_in_vma() calculates it and uses it to return true/false.
Return the address instead, allowing memory-failure to skip the call
to page_address_in_vma().

Acked-by: Miaohe Lin <linmiaohe@huawei.com>
Reviewed-by: Jane Chu <jane.chu@oracle.com>
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 include/linux/rmap.h |  2 +-
 mm/memory-failure.c  | 22 +++++++++++++---------
 mm/page_vma_mapped.c | 16 +++++++++-------
 3 files changed, 23 insertions(+), 17 deletions(-)

diff --git a/include/linux/rmap.h b/include/linux/rmap.h
index 0f906dc6d280..7229b9baf20d 100644
--- a/include/linux/rmap.h
+++ b/include/linux/rmap.h
@@ -730,7 +730,7 @@ int pfn_mkclean_range(unsigned long pfn, unsigned long nr_pages, pgoff_t pgoff,
 
 void remove_migration_ptes(struct folio *src, struct folio *dst, bool locked);
 
-int page_mapped_in_vma(struct page *page, struct vm_area_struct *vma);
+unsigned long page_mapped_in_vma(struct page *page, struct vm_area_struct *vma);
 
 /*
  * rmap_walk_control: To control rmap traversing for specific needs
diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index aec407788df1..0ad6b8936512 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -467,10 +467,11 @@ static void __add_to_kill(struct task_struct *tsk, struct page *p,
 }
 
 static void add_to_kill_anon_file(struct task_struct *tsk, struct page *p,
-				  struct vm_area_struct *vma,
-				  struct list_head *to_kill)
+		struct vm_area_struct *vma, struct list_head *to_kill,
+		unsigned long addr)
 {
-	unsigned long addr = page_address_in_vma(p, vma);
+	if (addr == -EFAULT)
+		return;
 	__add_to_kill(tsk, p, vma, to_kill, addr);
 }
 
@@ -595,7 +596,6 @@ struct task_struct *task_early_kill(struct task_struct *tsk, int force_early)
 static void collect_procs_anon(struct folio *folio, struct page *page,
 		struct list_head *to_kill, int force_early)
 {
-	struct vm_area_struct *vma;
 	struct task_struct *tsk;
 	struct anon_vma *av;
 	pgoff_t pgoff;
@@ -607,8 +607,10 @@ static void collect_procs_anon(struct folio *folio, struct page *page,
 	pgoff = page_to_pgoff(page);
 	rcu_read_lock();
 	for_each_process(tsk) {
+		struct vm_area_struct *vma;
 		struct anon_vma_chain *vmac;
 		struct task_struct *t = task_early_kill(tsk, force_early);
+		unsigned long addr;
 
 		if (!t)
 			continue;
@@ -617,9 +619,8 @@ static void collect_procs_anon(struct folio *folio, struct page *page,
 			vma = vmac->vma;
 			if (vma->vm_mm != t->mm)
 				continue;
-			if (!page_mapped_in_vma(page, vma))
-				continue;
-			add_to_kill_anon_file(t, page, vma, to_kill);
+			addr = page_mapped_in_vma(page, vma);
+			add_to_kill_anon_file(t, page, vma, to_kill, addr);
 		}
 	}
 	rcu_read_unlock();
@@ -642,6 +643,7 @@ static void collect_procs_file(struct folio *folio, struct page *page,
 	pgoff = page_to_pgoff(page);
 	for_each_process(tsk) {
 		struct task_struct *t = task_early_kill(tsk, force_early);
+		unsigned long addr;
 
 		if (!t)
 			continue;
@@ -654,8 +656,10 @@ static void collect_procs_file(struct folio *folio, struct page *page,
 			 * Assume applications who requested early kill want
 			 * to be informed of all such data corruptions.
 			 */
-			if (vma->vm_mm == t->mm)
-				add_to_kill_anon_file(t, page, vma, to_kill);
+			if (vma->vm_mm != t->mm)
+				continue;
+			addr = page_address_in_vma(page, vma);
+			add_to_kill_anon_file(t, page, vma, to_kill, addr);
 		}
 	}
 	rcu_read_unlock();
diff --git a/mm/page_vma_mapped.c b/mm/page_vma_mapped.c
index 53b8868ede61..c202eab84936 100644
--- a/mm/page_vma_mapped.c
+++ b/mm/page_vma_mapped.c
@@ -319,11 +319,12 @@ bool page_vma_mapped_walk(struct page_vma_mapped_walk *pvmw)
  * @page: the page to test
  * @vma: the VMA to test
  *
- * Returns 1 if the page is mapped into the page tables of the VMA, 0
- * if the page is not mapped into the page tables of this VMA.  Only
- * valid for normal file or anonymous VMAs.
+ * Return: The address the page is mapped at if the page is in the range
+ * covered by the VMA and present in the page table.  If the page is
+ * outside the VMA or not present, returns -EFAULT.
+ * Only valid for normal file or anonymous VMAs.
  */
-int page_mapped_in_vma(struct page *page, struct vm_area_struct *vma)
+unsigned long page_mapped_in_vma(struct page *page, struct vm_area_struct *vma)
 {
 	struct folio *folio = page_folio(page);
 	pgoff_t pgoff = folio->index + folio_page_idx(folio, page);
@@ -336,9 +337,10 @@ int page_mapped_in_vma(struct page *page, struct vm_area_struct *vma)
 
 	pvmw.address = vma_address(vma, pgoff, 1);
 	if (pvmw.address == -EFAULT)
-		return 0;
+		goto out;
 	if (!page_vma_mapped_walk(&pvmw))
-		return 0;
+		return -EFAULT;
 	page_vma_mapped_walk_done(&pvmw);
-	return 1;
+out:
+	return pvmw.address;
 }
-- 
2.43.0

[PATCH v3 04/11] mm: Make page_mapped_in_vma conditional on CONFIG_MEMORY_FAILURE

[Matthew Wilcox (Oracle)]

This function is only currently used by the memory-failure code, so
we can omit it if we're not compiling in the memory-failure code.

Suggested-by: Miaohe Lin <linmiaohe@huawei.com>
Reviewed-by: Jane Chu <jane.chu@oracle.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/page_vma_mapped.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/mm/page_vma_mapped.c b/mm/page_vma_mapped.c
index c202eab84936..ae5cc42aa208 100644
--- a/mm/page_vma_mapped.c
+++ b/mm/page_vma_mapped.c
@@ -314,6 +314,7 @@ bool page_vma_mapped_walk(struct page_vma_mapped_walk *pvmw)
 	return false;
 }
 
+#ifdef CONFIG_MEMORY_FAILURE
 /**
  * page_mapped_in_vma - check whether a page is really mapped in a VMA
  * @page: the page to test
@@ -344,3 +345,4 @@ unsigned long page_mapped_in_vma(struct page *page, struct vm_area_struct *vma)
 out:
 	return pvmw.address;
 }
+#endif
-- 
2.43.0

Re: [PATCH v3 04/11] mm: Make page_mapped_in_vma conditional on CONFIG_MEMORY_FAILURE

[Miaohe Lin]

On 2024/4/13 3:35, Matthew Wilcox (Oracle) wrote:
> This function is only currently used by the memory-failure code, so
> we can omit it if we're not compiling in the memory-failure code.
> 
> Suggested-by: Miaohe Lin <linmiaohe@huawei.com>
> Reviewed-by: Jane Chu <jane.chu@oracle.com>
> Reviewed-by: Oscar Salvador <osalvador@suse.de>
> Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>

Acked-by: Miaohe Lin <linmiaohe@huawei.com>
Thanks.
.

> ---
>  mm/page_vma_mapped.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/mm/page_vma_mapped.c b/mm/page_vma_mapped.c
> index c202eab84936..ae5cc42aa208 100644
> --- a/mm/page_vma_mapped.c
> +++ b/mm/page_vma_mapped.c
> @@ -314,6 +314,7 @@ bool page_vma_mapped_walk(struct page_vma_mapped_walk *pvmw)
>  	return false;
>  }
>  
> +#ifdef CONFIG_MEMORY_FAILURE
>  /**
>   * page_mapped_in_vma - check whether a page is really mapped in a VMA
>   * @page: the page to test
> @@ -344,3 +345,4 @@ unsigned long page_mapped_in_vma(struct page *page, struct vm_area_struct *vma)
>  out:
>  	return pvmw.address;
>  }
> +#endif
> 

[PATCH v3 05/11] mm/memory-failure: Convert shake_page() to shake_folio()

[Matthew Wilcox (Oracle)]

Removes two calls to compound_head().  Move the prototype to
internal.h; we definitely don't want code outside mm using it.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Jane Chu <jane.chu@oracle.com>
---
 include/linux/mm.h   |  1 -
 mm/hwpoison-inject.c | 11 ++++++-----
 mm/internal.h        |  1 +
 mm/memory-failure.c  | 15 ++++++++++-----
 4 files changed, 17 insertions(+), 11 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 326a4ce0cff8..b636504c176e 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -4028,7 +4028,6 @@ int mf_dax_kill_procs(struct address_space *mapping, pgoff_t index,
 extern int memory_failure(unsigned long pfn, int flags);
 extern void memory_failure_queue_kick(int cpu);
 extern int unpoison_memory(unsigned long pfn);
-extern void shake_page(struct page *p);
 extern atomic_long_t num_poisoned_pages __read_mostly;
 extern int soft_offline_page(unsigned long pfn, int flags);
 #ifdef CONFIG_MEMORY_FAILURE
diff --git a/mm/hwpoison-inject.c b/mm/hwpoison-inject.c
index d0548e382b6b..c9d653f51e45 100644
--- a/mm/hwpoison-inject.c
+++ b/mm/hwpoison-inject.c
@@ -15,7 +15,7 @@ static int hwpoison_inject(void *data, u64 val)
 {
 	unsigned long pfn = val;
 	struct page *p;
-	struct page *hpage;
+	struct folio *folio;
 	int err;
 
 	if (!capable(CAP_SYS_ADMIN))
@@ -25,16 +25,17 @@ static int hwpoison_inject(void *data, u64 val)
 		return -ENXIO;
 
 	p = pfn_to_page(pfn);
-	hpage = compound_head(p);
+	folio = page_folio(p);
 
 	if (!hwpoison_filter_enable)
 		goto inject;
 
-	shake_page(hpage);
+	shake_folio(folio);
 	/*
 	 * This implies unable to support non-LRU pages except free page.
 	 */
-	if (!PageLRU(hpage) && !PageHuge(p) && !is_free_buddy_page(p))
+	if (!folio_test_lru(folio) && !folio_test_hugetlb(folio) &&
+	    !is_free_buddy_page(p))
 		return 0;
 
 	/*
@@ -42,7 +43,7 @@ static int hwpoison_inject(void *data, u64 val)
 	 * the targeted owner (or on a free page).
 	 * memory_failure() will redo the check reliably inside page lock.
 	 */
-	err = hwpoison_filter(hpage);
+	err = hwpoison_filter(&folio->page);
 	if (err)
 		return 0;
 
diff --git a/mm/internal.h b/mm/internal.h
index ab8fcdeaf6eb..b20c736e05f2 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -1044,6 +1044,7 @@ static inline int find_next_best_node(int node, nodemask_t *used_node_mask)
 /*
  * mm/memory-failure.c
  */
+void shake_folio(struct folio *folio);
 extern int hwpoison_filter(struct page *p);
 
 extern u32 hwpoison_filter_dev_major;
diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 0ad6b8936512..bf9da2b46426 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -363,20 +363,25 @@ static int kill_proc(struct to_kill *tk, unsigned long pfn, int flags)
  * Unknown page type encountered. Try to check whether it can turn PageLRU by
  * lru_add_drain_all.
  */
-void shake_page(struct page *p)
+void shake_folio(struct folio *folio)
 {
-	if (PageHuge(p))
+	if (folio_test_hugetlb(folio))
 		return;
 	/*
 	 * TODO: Could shrink slab caches here if a lightweight range-based
 	 * shrinker will be available.
 	 */
-	if (PageSlab(p))
+	if (folio_test_slab(folio))
 		return;
 
 	lru_add_drain_all();
 }
-EXPORT_SYMBOL_GPL(shake_page);
+EXPORT_SYMBOL_GPL(shake_folio);
+
+static void shake_page(struct page *page)
+{
+	shake_folio(page_folio(page));
+}
 
 static unsigned long dev_pagemap_mapping_shift(struct vm_area_struct *vma,
 		unsigned long address)
@@ -1633,7 +1638,7 @@ static bool hwpoison_user_mappings(struct page *p, unsigned long pfn,
 	 * shake_page() again to ensure that it's flushed.
 	 */
 	if (mlocked)
-		shake_page(hpage);
+		shake_folio(folio);
 
 	/*
 	 * Now that the dirty bit has been propagated to the
-- 
2.43.0

Re: [PATCH v3 05/11] mm/memory-failure: Convert shake_page() to shake_folio()

[Miaohe Lin]

On 2024/4/13 3:35, Matthew Wilcox (Oracle) wrote:
> Removes two calls to compound_head().  Move the prototype to
> internal.h; we definitely don't want code outside mm using it.
> 
> Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
> Reviewed-by: Jane Chu <jane.chu@oracle.com>

Acked-by: Miaohe Lin <linmiaohe@huawei.com>
Thanks.
.

[PATCH v3 06/11] mm: Convert hugetlb_page_mapping_lock_write to folio

[Matthew Wilcox (Oracle)]

The page is only used to get the mapping, so the folio will do just
as well.  Both callers already have a folio available, so this saves
a call to compound_head().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Jane Chu  <jane.chu@oracle.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Acked-by: Miaohe Lin <linmiaohe@huawei.com>
---
 include/linux/hugetlb.h | 6 +++---
 mm/hugetlb.c            | 6 +++---
 mm/memory-failure.c     | 2 +-
 mm/migrate.c            | 2 +-
 4 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h
index 8c97ac48ae6d..aaa11eeb939d 100644
--- a/include/linux/hugetlb.h
+++ b/include/linux/hugetlb.h
@@ -178,7 +178,7 @@ bool hugetlbfs_pagecache_present(struct hstate *h,
 				 struct vm_area_struct *vma,
 				 unsigned long address);
 
-struct address_space *hugetlb_page_mapping_lock_write(struct page *hpage);
+struct address_space *hugetlb_folio_mapping_lock_write(struct folio *folio);
 
 extern int sysctl_hugetlb_shm_group;
 extern struct list_head huge_boot_pages[MAX_NUMNODES];
@@ -297,8 +297,8 @@ static inline unsigned long hugetlb_total_pages(void)
 	return 0;
 }
 
-static inline struct address_space *hugetlb_page_mapping_lock_write(
-							struct page *hpage)
+static inline struct address_space *hugetlb_folio_mapping_lock_write(
+							struct folio *folio)
 {
 	return NULL;
 }
diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index a8536349de13..001016993cca 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -2155,13 +2155,13 @@ static bool prep_compound_gigantic_folio_for_demote(struct folio *folio,
 /*
  * Find and lock address space (mapping) in write mode.
  *
- * Upon entry, the page is locked which means that page_mapping() is
+ * Upon entry, the folio is locked which means that folio_mapping() is
  * stable.  Due to locking order, we can only trylock_write.  If we can
  * not get the lock, simply return NULL to caller.
  */
-struct address_space *hugetlb_page_mapping_lock_write(struct page *hpage)
+struct address_space *hugetlb_folio_mapping_lock_write(struct folio *folio)
 {
-	struct address_space *mapping = page_mapping(hpage);
+	struct address_space *mapping = folio_mapping(folio);
 
 	if (!mapping)
 		return mapping;
diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index bf9da2b46426..52feaa9de9b4 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -1618,7 +1618,7 @@ static bool hwpoison_user_mappings(struct page *p, unsigned long pfn,
 		 * TTU_RMAP_LOCKED to indicate we have taken the lock
 		 * at this higher level.
 		 */
-		mapping = hugetlb_page_mapping_lock_write(hpage);
+		mapping = hugetlb_folio_mapping_lock_write(folio);
 		if (mapping) {
 			try_to_unmap(folio, ttu|TTU_RMAP_LOCKED);
 			i_mmap_unlock_write(mapping);
diff --git a/mm/migrate.c b/mm/migrate.c
index d87ce32645d4..fb0c83438fa8 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -1425,7 +1425,7 @@ static int unmap_and_move_huge_page(new_folio_t get_new_folio,
 			 * semaphore in write mode here and set TTU_RMAP_LOCKED
 			 * to let lower levels know we have taken the lock.
 			 */
-			mapping = hugetlb_page_mapping_lock_write(&src->page);
+			mapping = hugetlb_folio_mapping_lock_write(src);
 			if (unlikely(!mapping))
 				goto unlock_put_anon;
 
-- 
2.43.0

[PATCH v3 07/11] mm/memory-failure: Convert memory_failure() to use a folio

[Matthew Wilcox (Oracle)]

Saves dozens of calls to compound_head().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/memory-failure.c | 40 +++++++++++++++++++++-------------------
 1 file changed, 21 insertions(+), 19 deletions(-)

diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 52feaa9de9b4..2c12bcc9c550 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -2183,7 +2183,7 @@ static int memory_failure_dev_pagemap(unsigned long pfn, int flags,
 int memory_failure(unsigned long pfn, int flags)
 {
 	struct page *p;
-	struct page *hpage;
+	struct folio *folio;
 	struct dev_pagemap *pgmap;
 	int res = 0;
 	unsigned long page_flags;
@@ -2271,8 +2271,8 @@ int memory_failure(unsigned long pfn, int flags)
 		}
 	}
 
-	hpage = compound_head(p);
-	if (PageTransHuge(hpage)) {
+	folio = page_folio(p);
+	if (folio_test_large(folio)) {
 		/*
 		 * The flag must be set after the refcount is bumped
 		 * otherwise it may race with THP split.
@@ -2286,12 +2286,13 @@ int memory_failure(unsigned long pfn, int flags)
 		 * or unhandlable page.  The refcount is bumped iff the
 		 * page is a valid handlable page.
 		 */
-		SetPageHasHWPoisoned(hpage);
+		folio_set_has_hwpoisoned(folio);
 		if (try_to_split_thp_page(p) < 0) {
 			res = action_result(pfn, MF_MSG_UNSPLIT_THP, MF_IGNORED);
 			goto unlock_mutex;
 		}
 		VM_BUG_ON_PAGE(!page_count(p), p);
+		folio = page_folio(p);
 	}
 
 	/*
@@ -2302,9 +2303,9 @@ int memory_failure(unsigned long pfn, int flags)
 	 * The check (unnecessarily) ignores LRU pages being isolated and
 	 * walked by the page reclaim code, however that's not a big loss.
 	 */
-	shake_page(p);
+	shake_folio(folio);
 
-	lock_page(p);
+	folio_lock(folio);
 
 	/*
 	 * We're only intended to deal with the non-Compound page here.
@@ -2312,11 +2313,11 @@ int memory_failure(unsigned long pfn, int flags)
 	 * race window. If this happens, we could try again to hopefully
 	 * handle the page next round.
 	 */
-	if (PageCompound(p)) {
+	if (folio_test_large(folio)) {
 		if (retry) {
 			ClearPageHWPoison(p);
-			unlock_page(p);
-			put_page(p);
+			folio_unlock(folio);
+			folio_put(folio);
 			flags &= ~MF_COUNT_INCREASED;
 			retry = false;
 			goto try_again;
@@ -2332,29 +2333,29 @@ int memory_failure(unsigned long pfn, int flags)
 	 * folio_remove_rmap_*() in try_to_unmap_one(). So to determine page
 	 * status correctly, we save a copy of the page flags at this time.
 	 */
-	page_flags = p->flags;
+	page_flags = folio->flags;
 
 	if (hwpoison_filter(p)) {
 		ClearPageHWPoison(p);
-		unlock_page(p);
-		put_page(p);
+		folio_unlock(folio);
+		folio_put(folio);
 		res = -EOPNOTSUPP;
 		goto unlock_mutex;
 	}
 
 	/*
-	 * __munlock_folio() may clear a writeback page's LRU flag without
-	 * page_lock. We need wait writeback completion for this page or it
-	 * may trigger vfs BUG while evict inode.
+	 * __munlock_folio() may clear a writeback folio's LRU flag without
+	 * the folio lock. We need to wait for writeback completion for this
+	 * folio or it may trigger a vfs BUG while evicting inode.
 	 */
-	if (!PageLRU(p) && !PageWriteback(p))
+	if (!folio_test_lru(folio) && !folio_test_writeback(folio))
 		goto identify_page_state;
 
 	/*
 	 * It's very difficult to mess with pages currently under IO
 	 * and in many cases impossible, so we just avoid it here.
 	 */
-	wait_on_page_writeback(p);
+	folio_wait_writeback(folio);
 
 	/*
 	 * Now take care of user space mappings.
@@ -2368,7 +2369,8 @@ int memory_failure(unsigned long pfn, int flags)
 	/*
 	 * Torn down by someone else?
 	 */
-	if (PageLRU(p) && !PageSwapCache(p) && p->mapping == NULL) {
+	if (folio_test_lru(folio) && !folio_test_swapcache(folio) &&
+	    folio->mapping == NULL) {
 		res = action_result(pfn, MF_MSG_TRUNCATED_LRU, MF_IGNORED);
 		goto unlock_page;
 	}
@@ -2378,7 +2380,7 @@ int memory_failure(unsigned long pfn, int flags)
 	mutex_unlock(&mf_mutex);
 	return res;
 unlock_page:
-	unlock_page(p);
+	folio_unlock(folio);
 unlock_mutex:
 	mutex_unlock(&mf_mutex);
 	return res;
-- 
2.43.0

Re: [PATCH v3 07/11] mm/memory-failure: Convert memory_failure() to use a folio

[Miaohe Lin]

On 2024/4/13 3:35, Matthew Wilcox (Oracle) wrote:
>  
>  	/*
>  	 * Now take care of user space mappings.
> @@ -2368,7 +2369,8 @@ int memory_failure(unsigned long pfn, int flags)
>  	/*
>  	 * Torn down by someone else?
>  	 */
> -	if (PageLRU(p) && !PageSwapCache(p) && p->mapping == NULL) {
> +	if (folio_test_lru(folio) && !folio_test_swapcache(folio) &&
> +	    folio->mapping == NULL) {
>  		res = action_result(pfn, MF_MSG_TRUNCATED_LRU, MF_IGNORED);
>  		goto unlock_page;
>  	}
> @@ -2378,7 +2380,7 @@ int memory_failure(unsigned long pfn, int flags)
>  	mutex_unlock(&mf_mutex);
>  	return res;
>  unlock_page:

This label might be replaced with unlock_folio: ? Anyway, this patch looks good to me.
Acked-by: Miaohe Lin <linmiaohe@huawei.com>
Thanks.
.

[PATCH v3 08/11] mm/memory-failure: Convert hwpoison_user_mappings to take a folio

[Matthew Wilcox (Oracle)]

Pass the folio from the callers, and use it throughout instead of hpage.
Saves dozens of calls to compound_head().

Acked-by: Miaohe Lin <linmiaohe@huawei.com>
Reviewed-by: Jane Chu <jane.chu@oracle.com>
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/memory-failure.c | 30 +++++++++++++++---------------
 1 file changed, 15 insertions(+), 15 deletions(-)

diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 2c12bcc9c550..0fcf749682ab 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -1553,24 +1553,24 @@ static int get_hwpoison_page(struct page *p, unsigned long flags)
  * Do all that is necessary to remove user space mappings. Unmap
  * the pages and send SIGBUS to the processes if the data was dirty.
  */
-static bool hwpoison_user_mappings(struct page *p, unsigned long pfn,
-				  int flags, struct page *hpage)
+static bool hwpoison_user_mappings(struct folio *folio, struct page *p,
+		unsigned long pfn, int flags)
 {
-	struct folio *folio = page_folio(hpage);
 	enum ttu_flags ttu = TTU_IGNORE_MLOCK | TTU_SYNC | TTU_HWPOISON;
 	struct address_space *mapping;
 	LIST_HEAD(tokill);
 	bool unmap_success;
 	int forcekill;
-	bool mlocked = PageMlocked(hpage);
+	bool mlocked = folio_test_mlocked(folio);
 
 	/*
 	 * Here we are interested only in user-mapped pages, so skip any
 	 * other types of pages.
 	 */
-	if (PageReserved(p) || PageSlab(p) || PageTable(p) || PageOffline(p))
+	if (folio_test_reserved(folio) || folio_test_slab(folio) ||
+	    folio_test_pgtable(folio) || folio_test_offline(folio))
 		return true;
-	if (!(PageLRU(hpage) || PageHuge(p)))
+	if (!(folio_test_lru(folio) || folio_test_hugetlb(folio)))
 		return true;
 
 	/*
@@ -1580,7 +1580,7 @@ static bool hwpoison_user_mappings(struct page *p, unsigned long pfn,
 	if (!page_mapped(p))
 		return true;
 
-	if (PageSwapCache(p)) {
+	if (folio_test_swapcache(folio)) {
 		pr_err("%#lx: keeping poisoned page in swap cache\n", pfn);
 		ttu &= ~TTU_HWPOISON;
 	}
@@ -1591,11 +1591,11 @@ static bool hwpoison_user_mappings(struct page *p, unsigned long pfn,
 	 * XXX: the dirty test could be racy: set_page_dirty() may not always
 	 * be called inside page lock (it's recommended but not enforced).
 	 */
-	mapping = page_mapping(hpage);
-	if (!(flags & MF_MUST_KILL) && !PageDirty(hpage) && mapping &&
+	mapping = folio_mapping(folio);
+	if (!(flags & MF_MUST_KILL) && !folio_test_dirty(folio) && mapping &&
 	    mapping_can_writeback(mapping)) {
-		if (page_mkclean(hpage)) {
-			SetPageDirty(hpage);
+		if (folio_mkclean(folio)) {
+			folio_set_dirty(folio);
 		} else {
 			ttu &= ~TTU_HWPOISON;
 			pr_info("%#lx: corrupted page was clean: dropped without side effects\n",
@@ -1610,7 +1610,7 @@ static bool hwpoison_user_mappings(struct page *p, unsigned long pfn,
 	 */
 	collect_procs(folio, p, &tokill, flags & MF_ACTION_REQUIRED);
 
-	if (PageHuge(hpage) && !PageAnon(hpage)) {
+	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
@@ -1650,7 +1650,7 @@ static bool hwpoison_user_mappings(struct page *p, unsigned long pfn,
 	 * use a more force-full uncatchable kill to prevent
 	 * any accesses to the poisoned memory.
 	 */
-	forcekill = PageDirty(hpage) || (flags & MF_MUST_KILL) ||
+	forcekill = folio_test_dirty(folio) || (flags & MF_MUST_KILL) ||
 		    !unmap_success;
 	kill_procs(&tokill, forcekill, !unmap_success, pfn, flags);
 
@@ -2094,7 +2094,7 @@ static int try_memory_failure_hugetlb(unsigned long pfn, int flags, int *hugetlb
 
 	page_flags = folio->flags;
 
-	if (!hwpoison_user_mappings(p, pfn, flags, &folio->page)) {
+	if (!hwpoison_user_mappings(folio, p, pfn, flags)) {
 		folio_unlock(folio);
 		return action_result(pfn, MF_MSG_UNMAP_FAILED, MF_IGNORED);
 	}
@@ -2361,7 +2361,7 @@ int memory_failure(unsigned long pfn, int flags)
 	 * Now take care of user space mappings.
 	 * Abort on fail: __filemap_remove_folio() assumes unmapped page.
 	 */
-	if (!hwpoison_user_mappings(p, pfn, flags, p)) {
+	if (!hwpoison_user_mappings(folio, p, pfn, flags)) {
 		res = action_result(pfn, MF_MSG_UNMAP_FAILED, MF_IGNORED);
 		goto unlock_page;
 	}
-- 
2.43.0

[PATCH v3 09/11] mm/memory-failure: Add some folio conversions to unpoison_memory

[Matthew Wilcox (Oracle)]

Some of these folio APIs didn't exist when the unpoison_memory()
conversion was done originally.

Acked-by: Miaohe Lin <linmiaohe@huawei.com>
Reviewed-by: Jane Chu <jane.chu@oracle.com>
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 mm/memory-failure.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 0fcf749682ab..e42c5f2179dc 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -2550,8 +2550,8 @@ int unpoison_memory(unsigned long pfn)
 		goto unlock_mutex;
 	}
 
-	if (folio_test_slab(folio) || PageTable(&folio->page) ||
-	    folio_test_reserved(folio) || PageOffline(&folio->page))
+	if (folio_test_slab(folio) || folio_test_pgtable(folio) ||
+	    folio_test_reserved(folio) || folio_test_offline(folio))
 		goto unlock_mutex;
 
 	/*
@@ -2572,7 +2572,7 @@ int unpoison_memory(unsigned long pfn)
 
 	ghp = get_hwpoison_page(p, MF_UNPOISON);
 	if (!ghp) {
-		if (PageHuge(p)) {
+		if (folio_test_hugetlb(folio)) {
 			huge = true;
 			count = folio_free_raw_hwp(folio, false);
 			if (count == 0)
@@ -2588,7 +2588,7 @@ int unpoison_memory(unsigned long pfn)
 					 pfn, &unpoison_rs);
 		}
 	} else {
-		if (PageHuge(p)) {
+		if (folio_test_hugetlb(folio)) {
 			huge = true;
 			count = folio_free_raw_hwp(folio, false);
 			if (count == 0) {
-- 
2.43.0

[PATCH v3 10/11] mm/memory-failure: Use folio functions throughout collect_procs()

[Matthew Wilcox (Oracle)]

Saves a couple of calls to compound_head().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Jane Chu <jane.chu@oracle.com>
---
 mm/memory-failure.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index e42c5f2179dc..a9fa5901b48c 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -722,9 +722,9 @@ static void collect_procs(struct folio *folio, struct page *page,
 {
 	if (!folio->mapping)
 		return;
-	if (unlikely(PageKsm(page)))
+	if (unlikely(folio_test_ksm(folio)))
 		collect_procs_ksm(page, tokill, force_early);
-	else if (PageAnon(page))
+	else if (folio_test_anon(folio))
 		collect_procs_anon(folio, page, tokill, force_early);
 	else
 		collect_procs_file(folio, page, tokill, force_early);
-- 
2.43.0

Re: [PATCH v3 10/11] mm/memory-failure: Use folio functions throughout collect_procs()

[Miaohe Lin]

On 2024/4/13 3:35, Matthew Wilcox (Oracle) wrote:
> Saves a couple of calls to compound_head().
> 
> Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
> Reviewed-by: Jane Chu <jane.chu@oracle.com>

Acked-by: Miaohe Lin <linmiaohe@huawei.com>
Thanks.
.

> ---
>  mm/memory-failure.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
> index e42c5f2179dc..a9fa5901b48c 100644
> --- a/mm/memory-failure.c
> +++ b/mm/memory-failure.c
> @@ -722,9 +722,9 @@ static void collect_procs(struct folio *folio, struct page *page,
>  {
>  	if (!folio->mapping)
>  		return;
> -	if (unlikely(PageKsm(page)))
> +	if (unlikely(folio_test_ksm(folio)))
>  		collect_procs_ksm(page, tokill, force_early);
> -	else if (PageAnon(page))
> +	else if (folio_test_anon(folio))
>  		collect_procs_anon(folio, page, tokill, force_early);
>  	else
>  		collect_procs_file(folio, page, tokill, force_early);
> 

[PATCH v3 11/11] mm/memory-failure: Pass the folio to collect_procs_ksm()

[Matthew Wilcox (Oracle)]

We've already calculated it, so pass it in instead of recalculating it
in collect_procs_ksm().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Jane Chu <jane.chu@oracle.com>
---
 include/linux/ksm.h | 14 +++-----------
 mm/ksm.c            |  5 ++---
 mm/memory-failure.c |  2 +-
 3 files changed, 6 insertions(+), 15 deletions(-)

diff --git a/include/linux/ksm.h b/include/linux/ksm.h
index 358803cfd4d5..52c63a9c5a9c 100644
--- a/include/linux/ksm.h
+++ b/include/linux/ksm.h
@@ -81,15 +81,9 @@ struct folio *ksm_might_need_to_copy(struct folio *folio,
 
 void rmap_walk_ksm(struct folio *folio, struct rmap_walk_control *rwc);
 void folio_migrate_ksm(struct folio *newfolio, struct folio *folio);
-
-#ifdef CONFIG_MEMORY_FAILURE
-void collect_procs_ksm(struct page *page, struct list_head *to_kill,
-		       int force_early);
-#endif
-
-#ifdef CONFIG_PROC_FS
+void collect_procs_ksm(struct folio *folio, struct page *page,
+		struct list_head *to_kill, int force_early);
 long ksm_process_profit(struct mm_struct *);
-#endif /* CONFIG_PROC_FS */
 
 #else  /* !CONFIG_KSM */
 
@@ -120,12 +114,10 @@ static inline void ksm_might_unmap_zero_page(struct mm_struct *mm, pte_t pte)
 {
 }
 
-#ifdef CONFIG_MEMORY_FAILURE
-static inline void collect_procs_ksm(struct page *page,
+static inline void collect_procs_ksm(struct folio *folio, struct page *page,
 				     struct list_head *to_kill, int force_early)
 {
 }
-#endif
 
 #ifdef CONFIG_MMU
 static inline int ksm_madvise(struct vm_area_struct *vma, unsigned long start,
diff --git a/mm/ksm.c b/mm/ksm.c
index 108a4d167824..0bdd4d8b4c17 100644
--- a/mm/ksm.c
+++ b/mm/ksm.c
@@ -3172,12 +3172,11 @@ void rmap_walk_ksm(struct folio *folio, struct rmap_walk_control *rwc)
 /*
  * Collect processes when the error hit an ksm page.
  */
-void collect_procs_ksm(struct page *page, struct list_head *to_kill,
-		       int force_early)
+void collect_procs_ksm(struct folio *folio, struct page *page,
+		struct list_head *to_kill, int force_early)
 {
 	struct ksm_stable_node *stable_node;
 	struct ksm_rmap_item *rmap_item;
-	struct folio *folio = page_folio(page);
 	struct vm_area_struct *vma;
 	struct task_struct *tsk;
 
diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index a9fa5901b48c..c7cce73333f6 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -723,7 +723,7 @@ static void collect_procs(struct folio *folio, struct page *page,
 	if (!folio->mapping)
 		return;
 	if (unlikely(folio_test_ksm(folio)))
-		collect_procs_ksm(page, tokill, force_early);
+		collect_procs_ksm(folio, page, tokill, force_early);
 	else if (folio_test_anon(folio))
 		collect_procs_anon(folio, page, tokill, force_early);
 	else
-- 
2.43.0

Re: [PATCH v3 11/11] mm/memory-failure: Pass the folio to collect_procs_ksm()

[Miaohe Lin]

On 2024/4/13 3:35, Matthew Wilcox (Oracle) wrote:
> We've already calculated it, so pass it in instead of recalculating it
> in collect_procs_ksm().
> 
> Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
> Reviewed-by: Jane Chu <jane.chu@oracle.com>
> ---
>  include/linux/ksm.h | 14 +++-----------
>  mm/ksm.c            |  5 ++---
>  mm/memory-failure.c |  2 +-
>  3 files changed, 6 insertions(+), 15 deletions(-)

Reviewed-by: Miaohe Lin <linmiaohe@huawei.com>
Thanks.
.