Re: [PATCHv2 1/2 ] {xfrm, sctp} Stick to software crc32 even if hardware is capable of that

[Vlad Yasevich <vyasevich@gmail.com>]

On 10/11/2013 03:05 AM, Fan Du wrote:
>
>
> On 2013年10月10日 21:11, Neil Horman wrote:
>> On Thu, Oct 10, 2013 at 01:51:36PM +0800, Fan Du wrote:
>>> igb/ixgbe have hardware sctp checksum support, when this feature is
>>> enabled
>>> and also IPsec is armed to protect sctp traffic, ugly things happened as
>>> xfrm_output checks CHECKSUM_PARTIAL to do check sum operation(sum
>>> every thing
>>> up and pack the 16bits result in the checksum field). The result is fail
>>> establishment of sctp communication.
>>>
>> Shouldn't this be fixed in the xfrm code then?  E.g. check the device
>> features
>> for SCTP checksum offloading and and skip the checksum during xfrm
>> output if its
>> available?
>>
>> Or am I missing something?
>> Neil
>>
>>
>
>
>  From 014276de0877f11d46e1704114a7d91f19221a63 Mon Sep 17 00:00:00 2001
> From: Fan Du <fan.du@windriver.com>
> Date: Fri, 11 Oct 2013 14:24:33 +0800
> Subject: [PATCH 1/2] {xfrm, sctp} Stick to software crc32 even if
> hardware is
>   capable of that
>
> igb/ixgbe have hardware sctp checksum support, when this feature is enabled
> and also IPsec is armed to protect sctp traffic, ugly things happened as
> xfrm_output checks CHECKSUM_PARTIAL to do check sum operation(sum every
> thing
> up and pack the 16bits result in the checksum field). The result is fail
> establishment of sctp communication.
>
> Signed-off-by: Fan Du <fan.du@windriver.com>
> ---
> v2:
>    Leave ip_summed as CHECKSUM_PARTIAL as before, the second patch will
> fix this.
>
> ---
>   net/sctp/output.c |   14 +++++++++++++-
>   1 file changed, 13 insertions(+), 1 deletion(-)
>
> diff --git a/net/sctp/output.c b/net/sctp/output.c
> index 0ac3a65..6de6402 100644
> --- a/net/sctp/output.c
> +++ b/net/sctp/output.c
> @@ -372,6 +372,16 @@ static void sctp_packet_set_owner_w(struct sk_buff
> *skb, struct sock *sk)
>       atomic_inc(&sk->sk_wmem_alloc);
>   }
>
> +static int is_xfrm_armed(struct dst_entry *dst)
> +{
> +#ifdef CONFIG_XFRM
> +    /* If dst->xfrm is valid, this skb needs to be transformed */
> +    return dst->xfrm != NULL;
> +#else
> +    return 0;
> +#endif
> +}
> +

I would really prefer to have an accessor function to dst->xfrm, but
I see that everyone codes it inside the #ifdef.  Gack.

>   /* All packets are sent to the network through this function from
>    * sctp_outq_tail().
>    *
> @@ -536,7 +546,9 @@ int sctp_packet_transmit(struct sctp_packet *packet)
>        * by CRC32-C as described in <draft-ietf-tsvwg-sctpcsum-02.txt>.
>        */
>       if (!sctp_checksum_disable) {
> -        if (!(dst->dev->features & NETIF_F_SCTP_CSUM)) {
> +        if ((!(dst->dev->features & NETIF_F_SCTP_CSUM)) ||
> +            is_xfrm_armed(dst)) {
> +
>               __u32 crc32 = sctp_start_cksum((__u8 *)sh, cksum_buf_len);
>
>               /* 3) Put the resultant value into the checksum field in the

Acked-by: Vlad Yasevich <vyasevich@gmail.com>

-vlad