mm/pagewalkers: ACTION_AGAIN if pte_offset_map_lock() fails

Simple walk_page_range() users should set ACTION_AGAIN to retry when
pte_offset_map_lock() fails.

No need to check pmd_trans_unstable(): that was precisely to avoid the
possiblity of calling pte_offset_map() on a racily removed or inserted THP
entry, but such cases are now safely handled inside it.  Likewise there is
no need to check pmd_none() or pmd_bad() before calling it.

Link: https://lkml.kernel.org/r/[email protected]
Signed-off-by: Hugh Dickins <[email protected]>
Reviewed-by: SeongJae Park <[email protected]> for mm/damon part
Cc: Alistair Popple <[email protected]>
Cc: Anshuman Khandual <[email protected]>
Cc: Axel Rasmussen <[email protected]>
Cc: Christophe Leroy <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: David Hildenbrand <[email protected]>
Cc: "Huang, Ying" <[email protected]>
Cc: Ira Weiny <[email protected]>
Cc: Jason Gunthorpe <[email protected]>
Cc: Kirill A. Shutemov <[email protected]>
Cc: Lorenzo Stoakes <[email protected]>
Cc: Matthew Wilcox <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Miaohe Lin <[email protected]>
Cc: Mike Kravetz <[email protected]>
Cc: Mike Rapoport (IBM) <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Naoya Horiguchi <[email protected]>
Cc: Pavel Tatashin <[email protected]>
Cc: Peter Xu <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Qi Zheng <[email protected]>
Cc: Ralph Campbell <[email protected]>
Cc: Ryan Roberts <[email protected]>
Cc: Song Liu <[email protected]>
Cc: Steven Price <[email protected]>
Cc: Suren Baghdasaryan <[email protected]>
Cc: Thomas Hellström <[email protected]>
Cc: Will Deacon <[email protected]>
Cc: Yang Shi <[email protected]>
Cc: Yu Zhao <[email protected]>
Cc: Zack Rusin <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>

5 files changed, 36 insertions, 28 deletions

diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
index 6259dd432eeb..0d63b6a0f0d8 100644
--- a/fs/proc/task_mmu.c
+++ b/fs/proc/task_mmu.c
@@ -631,14 +631,11 @@ static int smaps_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
 		goto out;
 	}
 
-	if (pmd_trans_unstable(pmd))
-		goto out;
-	/*
-	 * The mmap_lock held all the way back in m_start() is what
-	 * keeps khugepaged out of here and from collapsing things
-	 * in here.
-	 */
 	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
+	if (!pte) {
+		walk->action = ACTION_AGAIN;
+		return 0;
+	}
 	for (; addr != end; pte++, addr += PAGE_SIZE)
 		smaps_pte_entry(pte, addr, walk);
 	pte_unmap_unlock(pte - 1, ptl);
@@ -1191,10 +1188,11 @@ out:
 		return 0;
 	}
 
-	if (pmd_trans_unstable(pmd))
-		return 0;
-
 	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
+	if (!pte) {
+		walk->action = ACTION_AGAIN;
+		return 0;
+	}
 	for (; addr != end; pte++, addr += PAGE_SIZE) {
 		ptent = *pte;
 
@@ -1538,9 +1536,6 @@ static int pagemap_pmd_range(pmd_t *pmdp, unsigned long addr, unsigned long end,
 		spin_unlock(ptl);
 		return err;
 	}
-
-	if (pmd_trans_unstable(pmdp))
-		return 0;
 #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
 
 	/*
@@ -1548,6 +1543,10 @@ static int pagemap_pmd_range(pmd_t *pmdp, unsigned long addr, unsigned long end,
 	 * goes beyond vma->vm_end.
 	 */
 	orig_pte = pte = pte_offset_map_lock(walk->mm, pmdp, addr, &ptl);
+	if (!pte) {
+		walk->action = ACTION_AGAIN;
+		return err;
+	}
 	for (; addr < end; pte++, addr += PAGE_SIZE) {
 		pagemap_entry_t pme;
 
@@ -1887,11 +1886,12 @@ static int gather_pte_stats(pmd_t *pmd, unsigned long addr,
 		spin_unlock(ptl);
 		return 0;
 	}
-
-	if (pmd_trans_unstable(pmd))
-		return 0;
 #endif
 	orig_pte = pte = pte_offset_map_lock(walk->mm, pmd, addr, &ptl);
+	if (!pte) {
+		walk->action = ACTION_AGAIN;
+		return 0;
+	}
 	do {
 		struct page *page = can_gather_numa_stats(*pte, vma, addr);
 		if (!page)
diff --git a/mm/damon/vaddr.c b/mm/damon/vaddr.c
index 37994fb6120c..e814f66dfc2e 100644
--- a/mm/damon/vaddr.c
+++ b/mm/damon/vaddr.c
@@ -318,9 +318,11 @@ static int damon_mkold_pmd_entry(pmd_t *pmd, unsigned long addr,
 		spin_unlock(ptl);
 	}
 
-	if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd)))
-		return 0;
 	pte = pte_offset_map_lock(walk->mm, pmd, addr, &ptl);
+	if (!pte) {
+		walk->action = ACTION_AGAIN;
+		return 0;
+	}
 	if (!pte_present(*pte))
 		goto out;
 	damon_ptep_mkold(pte, walk->vma, addr);
@@ -464,9 +466,11 @@ huge_out:
 regular_page:
 #endif	/* CONFIG_TRANSPARENT_HUGEPAGE */
 
-	if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd)))
-		return -EINVAL;
 	pte = pte_offset_map_lock(walk->mm, pmd, addr, &ptl);
+	if (!pte) {
+		walk->action = ACTION_AGAIN;
+		return 0;
+	}
 	if (!pte_present(*pte))
 		goto out;
 	folio = damon_get_folio(pte_pfn(*pte));
diff --git a/mm/mempolicy.c b/mm/mempolicy.c
index f06ca8c18e62..0241bb64978b 100644
--- a/mm/mempolicy.c
+++ b/mm/mempolicy.c
@@ -514,10 +514,11 @@ static int queue_folios_pte_range(pmd_t *pmd, unsigned long addr,
 	if (ptl)
 		return queue_folios_pmd(pmd, ptl, addr, end, walk);
 
-	if (pmd_trans_unstable(pmd))
-		return 0;
-
 	mapped_pte = pte = pte_offset_map_lock(walk->mm, pmd, addr, &ptl);
+	if (!pte) {
+		walk->action = ACTION_AGAIN;
+		return 0;
+	}
 	for (; addr != end; pte++, addr += PAGE_SIZE) {
 		if (!pte_present(*pte))
 			continue;
diff --git a/mm/mincore.c b/mm/mincore.c
index 2d5be013a25a..f33f6a0b1ded 100644
--- a/mm/mincore.c
+++ b/mm/mincore.c
@@ -113,12 +113,11 @@ static int mincore_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
 		goto out;
 	}
 
-	if (pmd_trans_unstable(pmd)) {
-		__mincore_unmapped_range(addr, end, vma, vec);
-		goto out;
-	}
-
 	ptep = pte_offset_map_lock(walk->mm, pmd, addr, &ptl);
+	if (!ptep) {
+		walk->action = ACTION_AGAIN;
+		return 0;
+	}
 	for (; addr != end; ptep++, addr += PAGE_SIZE) {
 		pte_t pte = *ptep;
 
diff --git a/mm/mlock.c b/mm/mlock.c
index 40b43f8740df..9f2b1173b1b1 100644
--- a/mm/mlock.c
+++ b/mm/mlock.c
@@ -329,6 +329,10 @@ static int mlock_pte_range(pmd_t *pmd, unsigned long addr,
 	}
 
 	start_pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
+	if (!start_pte) {
+		walk->action = ACTION_AGAIN;
+		return 0;
+	}
 	for (pte = start_pte; addr != end; pte++, addr += PAGE_SIZE) {
 		if (!pte_present(*pte))
 			continue;