Re: [PATCH v2] kasan: remove kasan_find_vm_area() to prevent possible deadlock

[Yeoreum Yun <yeoreum.yun@arm.com>]

Hi Sebastian,

> On 2025-07-07 09:33:54 [+0100], Yeoreum Yun wrote:
> > Hi Sebastian,
> Hi,
>
> > > what is DEPT?
> >
> > Please check the below patchset:
> >   https://lore.kernel.org/all/20250519091826.19752-1-byungchul@sk.com/
>
> Thank you.
>
> Would lockdep see this if check_region_inline() would have something
> like (minus missing exports, just illustrate the idea)
>
> diff --git a/mm/kasan/generic.c b/mm/kasan/generic.c
> index d54e89f8c3e76..c74e8e0863723 100644
> --- a/mm/kasan/generic.c
> +++ b/mm/kasan/generic.c
> @@ -171,6 +171,11 @@ static __always_inline bool check_region_inline(const void *addr,
>  	if (unlikely(size == 0))
>  		return true;
>
> +	{
> +		struct vmap_node *vn = vn = &vmap_nodes[0];
> +		might_lock(vn->busy.lock);
> +	}
> +
>  	if (unlikely(addr + size < addr))
>  		return !kasan_report(addr, size, write, ret_ip);
>
>
> ? Just to understand if lockdep is missing something essential or if
> DEPT was simply enabled why this "bad" accessed occurred and was able to
> see the lock chain which otherwise stays invisible.

No. I think lockdep could print this error situation without this patch.
Since lockdep prints other cycle in test enviroment. I've attached the
DEPT report.

Thanks.

--
Sincerely,
Yeoreum Yun