mm/hugetlb: fix uffd wr-protection for CoW optimization path

This patch fixes an issue that a hugetlb uffd-wr-protected mapping can be
writable even with uffd-wp bit set.  It only happens with hugetlb private
mappings, when someone firstly wr-protects a missing pte (which will
install a pte marker), then a write to the same page without any prior
access to the page.

Userfaultfd-wp trap for hugetlb was implemented in hugetlb_fault() before
reaching hugetlb_wp() to avoid taking more locks that userfault won't
need.  However there's one CoW optimization path that can trigger
hugetlb_wp() inside hugetlb_no_page(), which will bypass the trap.

This patch skips hugetlb_wp() for CoW and retries the fault if uffd-wp bit
is detected.  The new path will only trigger in the CoW optimization path
because generic hugetlb_fault() (e.g.  when a present pte was
wr-protected) will resolve the uffd-wp bit already.  Also make sure
anonymous UNSHARE won't be affected and can still be resolved, IOW only
skip CoW not CoR.

This patch will be needed for v5.19+ hence copy stable.

[peterx@redhat.com: v2]
  Link: https://lkml.kernel.org/r/ZBzOqwF2wrHgBVZb@x1n
[peterx@redhat.com: v3]
  Link: https://lkml.kernel.org/r/20230324142620.2344140-1-peterx@redhat.com
Link: https://lkml.kernel.org/r/20230321191840.1897940-1-peterx@redhat.com
Fixes: 166f3ec ("mm/hugetlb: hook page faults for uffd write protection")
Signed-off-by: Peter Xu <peterx@redhat.com>
Reported-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Tested-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Axel Rasmussen <axelrasmussen@google.com>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: Nadav Amit <nadav.amit@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

Author: xzpeter

mm/hugetlb.c

@@ -5478,7 +5478,7 @@ static vm_fault_t hugetlb_wp(struct mm_struct *mm, struct vm_area_struct *vma,
 		       struct folio *pagecache_folio, spinlock_t *ptl)
 {
 	const bool unshare = flags & FAULT_FLAG_UNSHARE;
-	pte_t pte;
+	pte_t pte = huge_ptep_get(ptep);
 	struct hstate *h = hstate_vma(vma);
 	struct page *old_page;
 	struct folio *new_folio;
@@ -5487,6 +5487,17 @@ static vm_fault_t hugetlb_wp(struct mm_struct *mm, struct vm_area_struct *vma,
 	unsigned long haddr = address & huge_page_mask(h);
 	struct mmu_notifier_range range;
 
+	/*
+	 * Never handle CoW for uffd-wp protected pages.  It should be only
+	 * handled when the uffd-wp protection is removed.
+	 *
+	 * Note that only the CoW optimization path (in hugetlb_no_page())
+	 * can trigger this, because hugetlb_fault() will always resolve
+	 * uffd-wp bit first.
+	 */
+	if (!unshare && huge_pte_uffd_wp(pte))
+		return 0;
+
 	/*
 	 * hugetlb does not support FOLL_FORCE-style write faults that keep the
 	 * PTE mapped R/O such as maybe_mkwrite() would do.
@@ -5500,7 +5511,6 @@ static vm_fault_t hugetlb_wp(struct mm_struct *mm, struct vm_area_struct *vma,
 		return 0;
 	}
 
-	pte = huge_ptep_get(ptep);
 	old_page = pte_page(pte);
 
 	delayacct_wpcopy_start();