Re: Are "skb->data" physically continuous?

[Nick Patavalis (npat@inaccessnetworks.com)]

On Tue, Sep 16, 2003 at 06:22:17PM +0300, Nick Patavalis wrote:
> On Tue, Sep 16, 2003 at 08:09:54AM -0700, Bluesea wrote:
> > I've a question, does the Linux TCP/IP stack calculate
> > the pseudo-header checksum if skb->ip_summed =
> > CHECKSUM_HW, or leave it to the driver?
> 
> Assuming that you are referring to the transmit path, and judging from
> the impementation of the e100 driver, it seems that the kernel has
> performed no checksum calculation if ip_summed == CHECKSUM_HW. Either
> the hardware or the driver has to do it. In the case of e100, the
> driver does the pseudo-header checksum calculation and the hardware
> does the rest.
> 

While this is what happens in the e100 driver, the kernel *does* seem
to calculate the *pseudo-header checksum* even in the "skb->ip_summed
== CHECKSUM_HW" case. For TCP packets this is done in
"net/ipv4/tcp_ipv4.c:tcp_v4_send_check()"

     <reformated>

    /* This routine computes an IPv4 TCP checksum. */
    void tcp_v4_send_check(struct sock *sk, struct tcphdr *th, int len, 
                           struct sk_buff *skb)
    {
        if (skb->ip_summed == CHECKSUM_HW) {
            th->check = ~tcp_v4_check(th, len, sk->saddr, sk->daddr, 0);
            skb->csum = offsetof(struct tcphdr, check);
        } else {
            th->check = tcp_v4_check(th, len, sk->saddr, sk->daddr,
                                     csum_partial((char *)th, 
                                     th->doff<<2, skb->csum));
        }
    }

which is called via the "ipv4_specific" jumptable-structure
initialized in: "net/ipv4/tcp_ipv4.c:"

    struct tcp_func ipv4_specific = {
        ip_queue_xmit,
        tcp_v4_send_check,
        tcp_v4_rebuild_header,

Why the e100 driver recalculates the pseude-header checksum, I don't
know.

/npat

-- 
Hey, maybe I could apply for a saint-hood from the Pope.  Does
somebody know what his email-address is? I'm so nice it makes you
puke.
  -- Linus Torvalds
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html