linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* sk_buff problem
@ 2003-09-04  7:42 Zhang haofeng
  0 siblings, 0 replies; 2+ messages in thread
From: Zhang haofeng @ 2003-09-04  7:42 UTC (permalink / raw)
  To: linux-kernel

Dear all,
    Does somebody have some knowledge about /net/ipv6/sit.c?
    I try to modify sit.c to cater for my requirements: I want to use
IPv4+UDP+IPv6+IPv6 Data to encapsulate IPv6 packets, thus every IPv6 packet
will be able to pass through some IPv4 NATs. As you know, sit.c is to
encapsulate IPv6 packets in IPv4 payload, so it seems that I only need to
add UDP layer between IPv4 header and IPv6 header in the encapsulated
packet.
    But I got problem when I receive an IPv4 UDP packet. I wanna
de-encapsulate the packet and send UDP data (which is an IPv6 packet) to
IPv6 upper layer protocol use netif_rx( ).  But it seems that I can not set
the correspondent field in sk_buff struct correctly.
    In sit.c, line 396, when 6to4 interface receive a IPv4 packet:
*****************
    if (!pskb_may_pull(skb, sizeof(struct ipv6hdr)))
        goto out;
...
    skb->mac.raw= skb->nh.raw;
    skb->nh.raw= skb-> data;
...
    netif_rx(skb);

    So in my modified sit.c, I modify the source code to :
*****************
    if (!pskb_may_pull(skb, sizeof(struct ipv6hdr)))
        goto out;
...
    skb->mac.ethernet->h_proto= __constant_htons(0x86DD);
    skb->nh.raw= skb->data+sizeof(struct udphdr);
    skb->protocol = __constant_htons(0x86DD);
...
    netif_rx(skb);
...

    The result is, from the debug information, I can see the IPv6 header,
the IPv6 data is put in sk_buff structure correctly. But
after I cannot use netif_rx() to send this de-encapsulated packet to IPv6
core protocol stack correctly, while the ICMPv4 UDP port unreachable message
is sent to the sender. I am confused :(
    I really appreciate your any kind of help, thanks in advance.....

Zhanghaofeng
Sep 3rd, 2003




^ permalink raw reply	[flat|nested] 2+ messages in thread

* sk_buff problem
@ 2003-09-03  4:47 Zhang haofeng
  0 siblings, 0 replies; 2+ messages in thread
From: Zhang haofeng @ 2003-09-03  4:47 UTC (permalink / raw)
  To: linux-kernel

Dear all,
    Does somebody have some knowledge about /net/ipv6/sit.c?
    I try to modify sit.c to cater for my requirements: I want to use
IPv4+UDP+IPv6+IPv6 Data to encapsulate IPv6 packets, thus every IPv6 packet
will be able to pass through some IPv4 NATs. As you know, sit.c is to
encapsulate IPv6 packets in IPv4 payload, so it seems that I only need to
add UDP layer between IPv4 header and IPv6 header in the encapsulated
packet.
    But I got problem when I receive an IPv4 UDP packet. I wanna
de-encapsulate the packet and send UDP data (which is an IPv6 packet) to
IPv6 upper layer protocol use netif_rx( ).  But it seems that I can not set
the correspondent field in sk_buff struct correctly.
    In sit.c, line 396, when 6to4 interface receive a IPv4 packet:
*****************
    if (!pskb_may_pull(skb, sizeof(struct ipv6hdr)))
        goto out;
...
    skb->mac.raw= skb->nh.raw;
    skb->nh.raw= skb-> data;
...
    netif_rx(skb);

    So in my modified sit.c, I modify the source code to :
*****************
    if (!pskb_may_pull(skb, sizeof(struct ipv6hdr)))
        goto out;
...
    skb->mac.raw= skb->nh.raw;
    skb->nh.raw= skb->data+sizeof(struct udphdr);
    skb->protocol = __constant_htons(0x86DD);
...
    netif_rx(skb);
...

    But it seems that I can not send the IPv6 packet to IPv6 upper layer,
while the ICMPv4 UDP port unreachable message is sent to the sender. I am
confused :(
    Any kind of  help is appreciated.

Zhanghaofeng
Sep 3rd, 2003



^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2003-09-04  7:45 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-09-04  7:42 sk_buff problem Zhang haofeng
  -- strict thread matches above, loose matches on Subject: below --
2003-09-03  4:47 Zhang haofeng

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).