thp: keep highpte mapped until it is no longer needed

Two users reported THP-related crashes on 32-bit x86 machines.  Their oops
reports indicated an invalid pte, and subsequent code inspection showed
that the highpte is actually used after unmap.

The fix is to unmap the pte only after all operations against it are
finished.

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reported-by: Ilya Dryomov <idryomov@gmail.com>
Reported-by: werner <w.landgraf@ru.ru>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Tested-by: Ilya Dryomov <idryomov@gmail.com>
Tested-by: Steven Rostedt <rostedt@goodmis.org
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

1 files changed, 2 insertions, 1 deletions

diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 004c9c2aac78..c4f634b3a48e 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -1837,9 +1837,9 @@ static void collapse_huge_page(struct mm_struct *mm,
         spin_lock(ptl);
         isolated = __collapse_huge_page_isolate(vma, address, pte);
         spin_unlock(ptl);
-        pte_unmap(pte);
 
         if (unlikely(!isolated)) {
+                pte_unmap(pte);
                 spin_lock(&mm->page_table_lock);
                 BUG_ON(!pmd_none(*pmd));
                 set_pmd_at(mm, address, pmd, _pmd);
@@ -1856,6 +1856,7 @@ static void collapse_huge_page(struct mm_struct *mm,
         anon_vma_unlock(vma->anon_vma);
 
         __collapse_huge_page_copy(pte, new_page, vma, address, ptl);
+        pte_unmap(pte);
         __SetPageUptodate(new_page);
         pgtable = pmd_pgtable(_pmd);
         VM_BUG_ON(page_count(pgtable) != 1);