Re: [PATCH] seccomp: fix the usage of get/put_seccomp_filter() in seccomp_get_filter()

[Kees Cook]

On Wed, Sep 20, 2017 at 5:56 AM, Oleg Nesterov <oleg@xxxxxxxxxx> wrote:
> As Chris explains, get_seccomp_filter() and put_seccomp_filter() can
> use the different filters, once we drop ->siglock task->seccomp.filter
> can be replaced by SECCOMP_FILTER_FLAG_TSYNC.
>
> Fixes: f8e529ed941b ("seccomp, ptrace: add support for dumping seccomp filters")
> Reported-by: Chris Salls <chrissalls5@xxxxxxxxx>
> Cc: stable@xxxxxxxxxxxxxxx
> Signed-off-by: Oleg Nesterov <oleg@xxxxxxxxxx>
> ---
>  kernel/seccomp.c | 14 +++++++++-----
>  1 file changed, 9 insertions(+), 5 deletions(-)
>
> diff --git a/kernel/seccomp.c b/kernel/seccomp.c
> index 98b59b5..897f153 100644
> --- a/kernel/seccomp.c
> +++ b/kernel/seccomp.c
> @@ -476,10 +476,8 @@ static inline void seccomp_filter_free(struct seccomp_filter *filter)
>         }
>  }
>
> -/* put_seccomp_filter - decrements the ref count of tsk->seccomp.filter */
> -void put_seccomp_filter(struct task_struct *tsk)
> +static void __put_seccomp_filter(struct seccomp_filter *orig)
>  {
> -       struct seccomp_filter *orig = tsk->seccomp.filter;
>         /* Clean up single-reference branches iteratively. */
>         while (orig && refcount_dec_and_test(&orig->usage)) {
>                 struct seccomp_filter *freeme = orig;
> @@ -488,6 +486,12 @@ void put_seccomp_filter(struct task_struct *tsk)
>         }
>  }
>
> +/* put_seccomp_filter - decrements the ref count of tsk->seccomp.filter */
> +void put_seccomp_filter(struct task_struct *tsk)
> +{
> +       __put_seccomp_filter(tsk->seccomp.filter);
> +}
> +
>  static void seccomp_init_siginfo(siginfo_t *info, int syscall, int reason)
>  {
>         memset(info, 0, sizeof(*info));
> @@ -908,13 +912,13 @@ long seccomp_get_filter(struct task_struct *task, unsigned long filter_off,
>         if (!data)
>                 goto out;
>
> -       get_seccomp_filter(task);
> +       refcount_inc(&filter->usage);
>         spin_unlock_irq(&task->sighand->siglock);
>
>         if (copy_to_user(data, fprog->filter, bpf_classic_proglen(fprog)))
>                 ret = -EFAULT;
>
> -       put_seccomp_filter(task);
> +       __put_seccomp_filter(filter);
>         return ret;

Given how reference counting is done for filters, I'd be happier with
leaving the get_seccomp_filter() as-is, and providing the
__put_seccomp_filter() as the only change here (i.e. don't open-code
the refcount_inc()).

-Kees

-- 
Kees Cook
Pixel Security