Re: [PATCH] fs: Handle register_shrinker failure

[Al Viro]

On Fri, Mar 24, 2017 at 09:55:40AM +0200, Nikolay Borisov wrote:
> register_shrinker allocates dynamic memory and thus is susceptible to failures
> under low-memory situation. Currently,get_userns ignores the return value of
> register_shrinker, potentially exposing not fully initialised object. This
> can lead to a NULL-ptr deref everytime shrinker->nr_deferred is referenced.
> 
> Fix this by failing to register the filesystem in case there is not enough
> memory to fully construct the shrinker object.
> 
> Signed-off-by: Nikolay Borisov <nborisov@xxxxxxxx>
> ---
>  fs/super.c | 14 +++++++++++++-
>  1 file changed, 13 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/super.c b/fs/super.c
> index b8b6a086c03b..964b18447c92 100644
> --- a/fs/super.c
> +++ b/fs/super.c
> @@ -518,7 +518,19 @@ struct super_block *sget_userns(struct file_system_type *type,
>  	hlist_add_head(&s->s_instances, &type->fs_supers);
>  	spin_unlock(&sb_lock);
>  	get_filesystem(type);
> -	register_shrinker(&s->s_shrink);
> +	err = register_shrinker(&s->s_shrink);
> +	if (err) {
> +		spin_lock(&sb_lock);
> +		list_del(&s->s_list);
> +		hlist_del(&s->s_instances);
> +		spin_unlock(&sb_lock);
> +
> +		up_write(&s->s_umount);
> +		destroy_super(s);
> +		put_filesystem(type);
> +		return ERR_PTR(err);

I really don't like that.  Your "remove it from all lists and pray that
nobody has picked a reference of any kind" at the very least needs a careful
written proof of correctness.  AFAICS, somebody might've found it on the
list and attempted to grab ->s_umount (grab_super() from another thread
calling sget()).  Then they'd block until your up_write() in there and
bugger the system up trying to play with ->s_umount in the object you've
freed.

NAK.  Yes, the bug is real, but this is not a solution.