mm/khugepaged: fix GUP-fast interaction by sending IPI

jira VULN-71586
cve-pre CVE-2025-38085
commit-author Jann Horn <[email protected]>
commit 2ba99c5
upstream-diff Resolved contextual conflicts due to lack of backported
  8d3c106

Since commit 70cbc3c ("mm: gup: fix the fast GUP race against THP
collapse"), the lockless_pages_from_mm() fastpath rechecks the pmd_t to
ensure that the page table was not removed by khugepaged in between.

However, lockless_pages_from_mm() still requires that the page table is
not concurrently freed.  Fix it by sending IPIs (if the architecture uses
semi-RCU-style page table freeing) before freeing/reusing page tables.

Link: https://lkml.kernel.org/r/[email protected]
Link: https://lkml.kernel.org/r/[email protected]
Link: https://lkml.kernel.org/r/[email protected]
Fixes: ba76149 ("thp: khugepaged")
	Signed-off-by: Jann Horn <[email protected]>
	Reviewed-by: Yang Shi <[email protected]>
	Acked-by: David Hildenbrand <[email protected]>
	Cc: John Hubbard <[email protected]>
	Cc: Peter Xu <[email protected]>
	Cc: <[email protected]>
	Signed-off-by: Andrew Morton <[email protected]>
(cherry picked from commit 2ba99c5)
	Signed-off-by: Marcin Wcisło <[email protected]>

Author: pvts-mat

include/asm-generic/tlb.h

@@ -222,12 +222,16 @@ extern void tlb_remove_table(struct mmu_gather *tlb, void *table);
 #define tlb_needs_table_invalidate() (true)
 #endif
 
+void tlb_remove_table_sync_one(void);
+
 #else
 
 #ifdef tlb_needs_table_invalidate
 #error tlb_needs_table_invalidate() requires MMU_GATHER_RCU_TABLE_FREE
 #endif
 
+static inline void tlb_remove_table_sync_one(void) { }
+
 #endif /* CONFIG_MMU_GATHER_RCU_TABLE_FREE */
 
 

mm/khugepaged.c

@@ -1138,6 +1138,7 @@ static void collapse_huge_page(struct mm_struct *mm,
 	_pmd = pmdp_collapse_flush(vma, address, pmd);
 	spin_unlock(pmd_ptl);
 	mmu_notifier_invalidate_range_end(&range);
+	tlb_remove_table_sync_one();
 
 	spin_lock(pte_ptl);
 	isolated = __collapse_huge_page_isolate(vma, address, pte,
@@ -1413,6 +1414,7 @@ static void collapse_and_free_pmd(struct mm_struct *mm, struct vm_area_struct *v
 	ptl = pmd_lock(vma->vm_mm, pmdp);
 	pmd = pmdp_collapse_flush(vma, addr, pmdp);
 	spin_unlock(ptl);
+	tlb_remove_table_sync_one();
 	mm_dec_nr_ptes(mm);
 	page_table_check_pte_clear_range(mm, addr, pmd);
 	pte_free(mm, pmd_pgtable(pmd));

mm/mmu_gather.c

@@ -140,7 +140,7 @@ static void tlb_remove_table_smp_sync(void *arg)
 	/* Simply deliver the interrupt */
 }
 
-static void tlb_remove_table_sync_one(void)
+void tlb_remove_table_sync_one(void)
 {
 	/*
 	 * This isn't an RCU grace period and hence the page-tables cannot be
@@ -164,8 +164,6 @@ static void tlb_remove_table_free(struct mmu_table_batch *batch)
 
 #else /* !CONFIG_MMU_GATHER_RCU_TABLE_FREE */
 
-static void tlb_remove_table_sync_one(void) { }
-
 static void tlb_remove_table_free(struct mmu_table_batch *batch)
 {
 	__tlb_remove_table_free(batch);