Re: [PATCH v2 1/1] xfrm: Use skb_mac_header_was_set() to check for MAC header presence

[Marcello Sylverster Bauer]

Hi Paolo,

On 9/5/23 11:59, Paolo Abeni wrote:
> On Mon, 2023-09-04 at 12:32 +0200, Marcello Sylvester Bauer wrote:
> > From: Marcello Sylvester Bauer <sylv@xxxxxxx>
> >
> > Add skb_mac_header_was_set() in xfrm4_remove_tunnel_encap() and
> > xfrm6_remove_tunnel_encap() to detect the presence of a MAC header.
> > This change prevents a kernel page fault on a non-zero mac_len when the
> > mac_header is not set.
> >
> > Signed-off-by: Marcello Sylvester Bauer <sylv@xxxxxxx>
>
> Please include a suitable fixes tag.
>
> Please also include in the commit message the stacktrace:
>
> https://lore.kernel.org/netdev/636d3434-d47a-4cd4-b3ba-7f7254317b64@xxxxxxx/
>
> trimming the asm code and lines starting with ' ? '

Sure, will be added to the next version in addition to a cover letter to 
give additional context to this patch.

> I think the real issue could be elsewhere, we should not reach here
> with mac_len > 0 && !skb_mac_header_was_set().
>
> Could you please try the following debug patch in your setup, and see
> if hints at some other relevant place?

Unfortunately the person with the hardware is OOO for two weeks. But I 
could ask him to test it when he gets back.

Thanks,
Marcello

> Thanks,
>
> Paolo
>
> ---
> diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
> index 4174c4b82d13..38ca2c7e50ca 100644
> --- a/include/linux/skbuff.h
> +++ b/include/linux/skbuff.h
> @@ -2793,6 +2793,7 @@ static inline void skb_reset_inner_headers(struct sk_buff *skb)
>   
>   static inline void skb_reset_mac_len(struct sk_buff *skb)
>   {
> +	WARN_ON_ONCE(!skb_mac_header_was_set(skb));
>   	skb->mac_len = skb->network_header - skb->mac_header;
>   }
>