From: Suren Baghdasaryan <surenb@xxxxxxxxxx> commit c137381f71aec755fbf47cd4e9bd4dce752c054c upstream. With recent changes necessitating mmap_lock to be held for write while expanding a stack, per-VMA locks should follow the same rules and be write-locked to prevent page faults into the VMA being expanded. Add the necessary locking. Cc: stable@xxxxxxxxxxxxxxx Signed-off-by: Suren Baghdasaryan <surenb@xxxxxxxxxx> Signed-off-by: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- mm/mmap.c | 4 ++++ 1 file changed, 4 insertions(+) --- a/mm/mmap.c +++ b/mm/mmap.c @@ -1975,6 +1975,8 @@ static int expand_upwards(struct vm_area return -ENOMEM; } + /* Lock the VMA before expanding to prevent concurrent page faults */ + vma_start_write(vma); /* * vma->vm_start/vm_end cannot change under us because the caller * is required to hold the mmap_lock in read mode. We need the @@ -2062,6 +2064,8 @@ int expand_downwards(struct vm_area_stru return -ENOMEM; } + /* Lock the VMA before expanding to prevent concurrent page faults */ + vma_start_write(vma); /* * vma->vm_start/vm_end cannot change under us because the caller * is required to hold the mmap_lock in read mode. We need the
