Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH v2] xfrm: Reset secpath in xfrm failure
To:     Myungho Jung <mhjungk@gmail.com>,
        Steffen Klassert <steffen.klassert@secunet.com>
Cc:     netdev@vger.kernel.org, linux-kernel@vger.kernel.org
References: <20190306215556.GA6104@myunghoj-Precision-5530>
From:   Eric Dumazet <eric.dumazet@gmail.com>
Message-ID: <1b3fc434-a5d2-286f-e7ef-027d165a5413@gmail.com>
Date:   Wed, 6 Mar 2019 14:31:25 -0800
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.9.1
MIME-Version: 1.0
In-Reply-To: <20190306215556.GA6104@myunghoj-Precision-5530>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


On 03/06/2019 01:55 PM, Myungho Jung wrote:
> In esp4_gro_receive() and esp6_gro_receive(), secpath can be allocated
> without adding xfrm state to xvec. Then, sp->xvec[sp->len - 1] would
> fail and result in dereferencing invalid pointer in esp4_gso_segment()
> and esp6_gso_segment(). Reset secpath if xfrm function returns error.
> 
> Fixes: 7785bba299a8 ("esp: Add a software GRO codepath")
> Reported-by: syzbot+b69368fd933c6c592f4c@syzkaller.appspotmail.com
> Signed-off-by: Myungho Jung <mhjungk@gmail.com>
> ---
> Changes in v2:
>   - Add fixes tag.
> 
>  net/ipv4/esp4_offload.c | 9 +++++++--
>  net/ipv6/esp6_offload.c | 9 +++++++--
>  2 files changed, 14 insertions(+), 4 deletions(-)
> 
> diff --git a/net/ipv4/esp4_offload.c b/net/ipv4/esp4_offload.c
> index 8756e0e790d2..7329e40c73f6 100644
> --- a/net/ipv4/esp4_offload.c
> +++ b/net/ipv4/esp4_offload.c
> @@ -51,14 +51,18 @@ static struct sk_buff *esp4_gro_receive(struct list_head *head,
>  		if (!sp)
>  			goto out;
>  
> -		if (sp->len == XFRM_MAX_DEPTH)
> +		if (sp->len == XFRM_MAX_DEPTH) {
> +			secpath_reset(skb);
>  			goto out;
> +		}
>  
>  		x = xfrm_state_lookup(dev_net(skb->dev), skb->mark,
>  				      (xfrm_address_t *)&ip_hdr(skb)->daddr,
>  				      spi, IPPROTO_ESP, AF_INET);
> -		if (!x)
> +		if (!x) {
> +			secpath_reset(skb);
>  			goto out;
> +		}
>  

I suggest another exit label, so that you replace "goto out" by "goto out_reset";

diff --git a/net/ipv4/esp4_offload.c b/net/ipv4/esp4_offload.c
index 8756e0e790d2a94a5b4a587c3bc3de0673baf2c4..76f754f6692696ba2aa8c9eb03b68b92d1e39ee1 100644
--- a/net/ipv4/esp4_offload.c
+++ b/net/ipv4/esp4_offload.c
@@ -82,6 +82,8 @@ static struct sk_buff *esp4_gro_receive(struct list_head *head,
        xfrm_input(skb, IPPROTO_ESP, spi, -2);
 
        return ERR_PTR(-EINPROGRESS);
+out_reset:
+       secpath_reset(skb);
 out:
        skb_push(skb, offset);
        NAPI_GRO_CB(skb)->same_flow = 0;


>  		sp->xvec[sp->len++] = x;
>  		sp->olen++;
> @@ -66,6 +70,7 @@ static struct sk_buff *esp4_gro_receive(struct list_head *head,
>  		xo = xfrm_offload(skb);
>  		if (!xo) {
>  			xfrm_state_put(x);
> +			secpath_reset(skb);
>  			goto out;
>  		}
>  	}