["PATCH 5.4.y"] userfaultfd: fix mmap_changing checking in mfill_atomic_hugetlb

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



From: Lokesh Gidra <lokeshgidra@xxxxxxxxxx>

In mfill_atomic_hugetlb(), mmap_changing isn't being checked
again if we drop mmap_lock and reacquire it. When the lock is not held,
mmap_changing could have been incremented. This is also inconsistent
with the behavior in mfill_atomic().

Link: https://lkml.kernel.org/r/20240117223729.1444522-1-lokeshgidra@xxxxxxxxxx
Fixes: df2cc96e77011 ("userfaultfd: prevent non-cooperative events vs mcopy_atomic races")
Signed-off-by: Lokesh Gidra <lokeshgidra@xxxxxxxxxx>
Cc: Andrea Arcangeli <aarcange@xxxxxxxxxx>
Cc: Mike Rapoport <rppt@xxxxxxxxxx>
Cc: Axel Rasmussen <axelrasmussen@xxxxxxxxxx>
Cc: Brian Geffon <bgeffon@xxxxxxxxxx>
Cc: David Hildenbrand <david@xxxxxxxxxx>
Cc: Jann Horn <jannh@xxxxxxxxxx>
Cc: Kalesh Singh <kaleshsingh@xxxxxxxxxx>
Cc: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
Cc: Nicolas Geoffray <ngeoffray@xxxxxxxxxx>
Cc: Peter Xu <peterx@xxxxxxxxxx>
Cc: Suren Baghdasaryan <surenb@xxxxxxxxxx>
Cc: <stable@xxxxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
(cherry picked from commit 67695f18d55924b2013534ef3bdc363bc9e14605)
Signed-off-by: Mike Rapoport (IBM) <rppt@xxxxxxxxxx>
---
 mm/userfaultfd.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/mm/userfaultfd.c b/mm/userfaultfd.c
index 6fa66e2111ea..e8758304a9f3 100644
--- a/mm/userfaultfd.c
+++ b/mm/userfaultfd.c
@@ -177,6 +177,7 @@ static __always_inline ssize_t __mcopy_atomic_hugetlb(struct mm_struct *dst_mm,
 					      unsigned long dst_start,
 					      unsigned long src_start,
 					      unsigned long len,
+					      bool *mmap_changing,
 					      bool zeropage)
 {
 	int vm_alloc_shared = dst_vma->vm_flags & VM_SHARED;
@@ -308,6 +309,15 @@ static __always_inline ssize_t __mcopy_atomic_hugetlb(struct mm_struct *dst_mm,
 				goto out;
 			}
 			down_read(&dst_mm->mmap_sem);
+			/*
+			 * If memory mappings are changing because of non-cooperative
+			 * operation (e.g. mremap) running in parallel, bail out and
+			 * request the user to retry later
+			 */
+			if (mmap_changing && READ_ONCE(*mmap_changing)) {
+				err = -EAGAIN;
+				break;
+			}
 
 			dst_vma = NULL;
 			goto retry;
@@ -389,6 +399,7 @@ extern ssize_t __mcopy_atomic_hugetlb(struct mm_struct *dst_mm,
 				      unsigned long dst_start,
 				      unsigned long src_start,
 				      unsigned long len,
+				      bool *mmap_changing,
 				      bool zeropage);
 #endif /* CONFIG_HUGETLB_PAGE */
 
@@ -506,7 +517,8 @@ static __always_inline ssize_t __mcopy_atomic(struct mm_struct *dst_mm,
 	 */
 	if (is_vm_hugetlb_page(dst_vma))
 		return  __mcopy_atomic_hugetlb(dst_mm, dst_vma, dst_start,
-						src_start, len, zeropage);
+					       src_start, len, mmap_changing,
+					       zeropage);
 
 	if (!vma_is_anonymous(dst_vma) && !vma_is_shmem(dst_vma))
 		goto out_unlock;
-- 
2.43.0





[Index of Archives]     [Linux Kernel]     [Kernel Development Newbies]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite Hiking]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux