[PATCH v2] mm: make page_mapped_in_vma() hugetlb walk aware

[Jane Chu <jane.chu@oracle.com>]

When a process consumes a UE in a page, the memory failure handler
attempts to collect information for a potential SIGBUS.
If the page is an anonymous page, page_mapped_in_vma(page, vma) is
invoked in order to
  1. retrieve the vaddr from the process' address space,
  2. verify that the vaddr is indeed mapped to the poisoned page,
where 'page' is the precise small page with UE.

It's been observed that when injecting poison to a non-head subpage
of an anonymous hugetlb page, no SIGBUS show up; while injecting to
the head page produces a SIGBUS. The casue is that, though hugetlb_walk()
returns a valid pmd entry (on x86), but check_pte() detects mismatch
between the head page per the pmd and the input subpage. Thus the vaddr
is considered not mapped to the subpage and the process is not collected
for SIGBUS purpose.  This is the calling stack
      collect_procs_anon
        page_mapped_in_vma
          page_vma_mapped_walk
            hugetlb_walk
              huge_pte_lock
                check_pte

check_pte() header says that it
"check if [pvmw->pfn, @pvmw->pfn + @pvmw->nr_pages) is mapped at the @pvmw->pte"
but practically works only if pvmw->pfn is the head page pfn at pvmw->pte.
Hindsight acknowledging that some pvmw->pte could point to a hugepage of
some sort such that it makes sense to make check_pte() work for hugepage.

Signed-off-by: Jane Chu <jane.chu@oracle.com>
---
 mm/page_vma_mapped.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/mm/page_vma_mapped.c b/mm/page_vma_mapped.c
index 81839a9e74f1..367209e65830 100644
--- a/mm/page_vma_mapped.c
+++ b/mm/page_vma_mapped.c
@@ -84,6 +84,7 @@ static bool map_pte(struct page_vma_mapped_walk *pvmw, pmd_t *pmdvalp,
  * mapped at the @pvmw->pte
  * @pvmw: page_vma_mapped_walk struct, includes a pair pte and pfn range
  * for checking
+ * @pte_nr: the number of small pages described by @pvmw->pte.
  *
  * page_vma_mapped_walk() found a place where pfn range is *potentially*
  * mapped. check_pte() has to validate this.
@@ -100,7 +101,7 @@ static bool map_pte(struct page_vma_mapped_walk *pvmw, pmd_t *pmdvalp,
  * Otherwise, return false.
  *
  */
-static bool check_pte(struct page_vma_mapped_walk *pvmw)
+static bool check_pte(struct page_vma_mapped_walk *pvmw, unsigned long pte_nr)
 {
 	unsigned long pfn;
 	pte_t ptent = ptep_get(pvmw->pte);
@@ -133,7 +134,11 @@ static bool check_pte(struct page_vma_mapped_walk *pvmw)
 		pfn = pte_pfn(ptent);
 	}
 
-	return (pfn - pvmw->pfn) < pvmw->nr_pages;
+	if ((pfn + pte_nr - 1) < pvmw->pfn)
+		return false;
+	if (pfn > (pvmw->pfn + pvmw->nr_pages - 1))
+		return false;
+	return true;
 }
 
 /* Returns true if the two ranges overlap.  Careful to not overflow. */
@@ -208,7 +213,7 @@ bool page_vma_mapped_walk(struct page_vma_mapped_walk *pvmw)
 			return false;
 
 		pvmw->ptl = huge_pte_lock(hstate, mm, pvmw->pte);
-		if (!check_pte(pvmw))
+		if (!check_pte(pvmw, pages_per_huge_page(hstate)))
 			return not_found(pvmw);
 		return true;
 	}
@@ -291,7 +296,7 @@ bool page_vma_mapped_walk(struct page_vma_mapped_walk *pvmw)
 			goto next_pte;
 		}
 this_pte:
-		if (check_pte(pvmw))
+		if (check_pte(pvmw, 1))
 			return true;
 next_pte:
 		do {
-- 
2.39.3