| Message ID | 20201209032155.564991-1-minchan@kernel.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | mm/madvise: remove racy mm ownership check | expand |
On Tue, Dec 8, 2020 at 7:22 PM Minchan Kim <minchan@kernel.org> wrote: > > Jann spotted the security hole due to race of mm ownership check. > If the task is sharing the mm_struct but goes through execve() > before mm_access(), it could skip process_madvise_behavior_valid > check. That makes *any advice hint* to reach into the remote process. > > This patch removes the mm ownership check. Applied directly - I'm not sure Andrew has anything else pending, so might as well short-circuit it. Linus
diff --git a/mm/madvise.c b/mm/madvise.c index a8d8d48a57fe..13f5677b9322 100644 --- a/mm/madvise.c +++ b/mm/madvise.c @@ -1204,8 +1204,7 @@ SYSCALL_DEFINE5(process_madvise, int, pidfd, const struct iovec __user *, vec, goto put_pid; } - if (task->mm != current->mm && - !process_madvise_behavior_valid(behavior)) { + if (!process_madvise_behavior_valid(behavior)) { ret = -EINVAL; goto release_task; }
Jann spotted the security hole due to race of mm ownership check. If the task is sharing the mm_struct but goes through execve() before mm_access(), it could skip process_madvise_behavior_valid check. That makes *any advice hint* to reach into the remote process. This patch removes the mm ownership check. With it, it will lose the ability that local process could give *any* advice hint with vector interface for some reason(e.g., performance). Since there is no concrete example in upstream yet, it would be better to remove the abiliity at this moment and need to review when such new advice comes up. Cc: security@kernel.org Fixes: ecb8ac8b1f14 ("mm/madvise: introduce process_madvise() syscall: an external memory hinting API") Reported-by: Jann Horn <jannh@google.com> Suggested-by: Jann Horn <jannh@google.com> Signed-off-by: Minchan Kim <minchan@kernel.org> --- mm/madvise.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)