Re: [PATCH] esp: fix page frag reference leak on skb_to_sgvec failure

[Steffen Klassert <steffen.klassert@secunet.com>]

On Sat, May 16, 2026 at 11:20:01AM +0200, Alessandro Schino wrote:
> From: e521588 <alessandro.schino@sbb.ch>
> 
> In esp_output_tail(), when esp->inplace is false, the old skb page frags
> are replaced with a new page from the xfrm page_frag cache. The source
> scatterlist (sg) is built from the old frags before the replacement, and
> esp_ssg_unref() is responsible for releasing the old page references
> after the crypto operation completes.
> 
> However, if the second skb_to_sgvec() call (which builds the destination
> scatterlist from the new page) fails, the code jumps to error_free which
> only calls kfree(tmp). The old page frag references captured in the
> source scatterlist are never released:
> 
>   1. sg[] is built from old frags via skb_to_sgvec() (no extra get_page)
>   2. nr_frags is set to 1 and frag[0] is replaced with the new page
>   3. Second skb_to_sgvec() fails -> goto error_free
>   4. kfree(tmp) frees the sg[] memory but old frags are not unref'd
>   5. kfree_skb() only releases frag[0] (the new page), not the old ones
> 
> Fix this by introducing esp_ssg_unref_frags(), a variant of
> esp_ssg_unref() that unconditionally unrefs the source scatterlist
> frags without checking req->src and req->dst, since those fields are
> not yet initialized at this point. The new function is called before
> the goto error_free when the second skb_to_sgvec() fails, at which
> point the source scatterlist in tmp is fully populated.
> 
> The same issue exists in both esp4 and esp6 as the code is identical.
> 
> Fixes: cac2661c53f3 ("esp4: Avoid skb_cow_data whenever possible")
> Fixes: 03e2a30f6a27 ("esp6: Avoid skb_cow_data whenever possible")
> 
> Signed-off-by: Alessandro Schino <7991aleschino@gmail.com>
> ---
>  net/ipv4/esp4.c | 22 +++++++++++++++++++++-
>  net/ipv6/esp6.c | 22 +++++++++++++++++++++-
>  2 files changed, 42 insertions(+), 2 deletions(-)
> 
> diff --git a/net/ipv4/esp4.c b/net/ipv4/esp4.c
> index 5c4b162c1bb5..bc5245c6a29a 100644
> --- a/net/ipv4/esp4.c
> +++ b/net/ipv4/esp4.c
> @@ -119,6 +119,26 @@ static void esp_ssg_unref(struct xfrm_state *x, void *tmp, struct sk_buff *skb)
>  				       skb->pp_recycle);
>  }
>  
> +static void esp_ssg_unref_frags(struct xfrm_state *x, void *tmp,
> +                                struct sk_buff *skb)
> +{
> +	struct crypto_aead *aead = x->data;
> +	int extralen = 0;
> +	u8 *iv;
> +	struct aead_request *req;
> +	struct scatterlist *sg;
> +
> +	if (x->props.flags & XFRM_STATE_ESN)
> +		extralen += sizeof(struct esp_output_extra);
> +
> +	iv = esp_tmp_iv(aead, tmp, extralen);
> +	req = esp_tmp_req(aead, iv);
> +
> +	for (sg = sg_next(req->src); sg; sg = sg_next(sg))
> +		skb_page_unref(page_to_netmem(sg_page(sg)),
> +			       skb->pp_recycle);
> +}

Sorry, I wasn't exactly precise. Instead of duplicating esp_ssg_unref()
it might make more sense to add a flag to esp_ssg_unref() and check
src and dst of the crypto request based on that flag.