| Message ID | CAG48ez1-PBCdv3y8pn-Ty-b+FmBSLwDuVKFSt8h7wARLy0dF-Q@mail.gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [1/4] mm/gup_benchmark: Take the mmap lock around GUP | expand |
On Tue, Sep 29, 2020 at 6:20 PM Jann Horn <jannh@google.com> wrote: > create_elf_tables() runs after setup_new_exec(), so other tasks can > already access our new mm and do things like process_madvise() on it. > (At the time I'm writing this commit, process_madvise() is not in mainline > yet, but has been in akpm's tree for some time.) > > While I believe that there are currently no APIs that would actually allow > another process to mess up our VMA tree (process_madvise() is limited to > MADV_COLD and MADV_PAGEOUT, and uring and userfaultfd cannot reach an mm > under which no syscalls have been executed yet), this seems like an > accident waiting to happen. > > Let's make sure that we always take the mmap lock around GUP paths as long > as another process might be able to see the mm. > > (Yes, this diff looks suspicious because we drop the lock before doing > anything with `vma`, but that's because we actually don't do anything with > it apart from the NULL check.) > > Signed-off-by: Jann Horn <jannh@google.com> Thanks for these cleanups :) Acked-by: Michel Lespinasse <walken@google.com>
diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index 40ec0b9b4b4f..cd7c574a91a4 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -309,7 +309,10 @@ create_elf_tables(struct linux_binprm *bprm, const struct elfhdr *exec, * Grow the stack manually; some architectures have a limit on how * far ahead a user-space access may be in order to grow the stack. */ + if (mmap_read_lock_killable(mm)) + return -EINTR; vma = find_extend_vma(mm, bprm->p); + mmap_read_unlock(mm); if (!vma) return -EFAULT;
create_elf_tables() runs after setup_new_exec(), so other tasks can already access our new mm and do things like process_madvise() on it. (At the time I'm writing this commit, process_madvise() is not in mainline yet, but has been in akpm's tree for some time.) While I believe that there are currently no APIs that would actually allow another process to mess up our VMA tree (process_madvise() is limited to MADV_COLD and MADV_PAGEOUT, and uring and userfaultfd cannot reach an mm under which no syscalls have been executed yet), this seems like an accident waiting to happen. Let's make sure that we always take the mmap lock around GUP paths as long as another process might be able to see the mm. (Yes, this diff looks suspicious because we drop the lock before doing anything with `vma`, but that's because we actually don't do anything with it apart from the NULL check.) Signed-off-by: Jann Horn <jannh@google.com> --- fs/binfmt_elf.c | 3 +++ 1 file changed, 3 insertions(+)