The userfaultfd hugetlb tests detect a resv_huge_pages underflow. This
happens when hugetlb_mcopy_atomic_pte() is called with !is_continue on
an index for which we already have a page in the cache. When this
happens, we allocate a second page, double consuming the reservation,
and then fail to insert the page into the cache and return -EEXIST.
To fix this, we first if there exists a page in the cache which already
consumed the reservation, and return -EEXIST immediately if so.
There is still a rare condition where we fail to copy the page contents
AND race with a call for hugetlb_no_page() for this index and again we
will underflow resv_huge_pages. That is fixed in a more complicated
patch not targeted for -stable.
Test:
Hacked the code locally such that resv_huge_pages underflows produce
a warning, then:
./tools/testing/selftests/vm/userfaultfd hugetlb_shared 10
2 /tmp/kokonut_test/huge/userfaultfd_test && echo test success
./tools/testing/selftests/vm/userfaultfd hugetlb 10
2 /tmp/kokonut_test/huge/userfaultfd_test && echo test success
Both tests succeed and produce no warnings. After the
test runs number of free/resv hugepages is correct.
Signed-off-by: Mina Almasry <almasrymina@xxxxxxxxxx>
Cc: Axel Rasmussen <axelrasmussen@xxxxxxxxxx>
Cc: Peter Xu <peterx@xxxxxxxxxx>
Cc: linux-mm@xxxxxxxxx
Cc: Mike Kravetz <mike.kravetz@xxxxxxxxxx>
Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Cc: linux-mm@xxxxxxxxx
Cc: linux-kernel@xxxxxxxxxxxxxxx
Cc: stable@xxxxxxxxxxxxxxx
---
mm/hugetlb.c | 14 ++++++++++++--
1 file changed, 12 insertions(+), 2 deletions(-)
diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index ead5d12e0604..76e2a6efc165 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -4925,10 +4925,20 @@ int hugetlb_mcopy_atomic_pte(struct mm_struct *dst_mm,
if (!page)
goto out;
} else if (!*pagep) {
- ret = -ENOMEM;
+ /* If a page already exists, then it's UFFDIO_COPY for
+ * a non-missing case. Return -EEXIST.
+ */
+ if (vm_shared &&
+ hugetlbfs_pagecache_present(h, dst_vma, dst_addr)) {
+ ret = -EEXIST;
+ goto out;
+ }
+
page = alloc_huge_page(dst_vma, dst_addr, 0);
- if (IS_ERR(page))
+ if (IS_ERR(page)) {
+ ret = -ENOMEM;
goto out;
+ }
ret = copy_huge_page_from_user(page,
(const void __user *) src_addr,
--
2.32.0.rc0.204.g9fa02ecfa5-goog
