On Tue, May 16, 2006 at 12:06:31AM -0700, David S. Miller ([EMAIL PROTECTED])
wrote:
> From: Evgeniy Polyakov <[EMAIL PROTECTED]>
> Date: Tue, 16 May 2006 10:59:23 +0400
>
> > And what if we use ESP which would place it's hashed sequence number as
> > port?
>
> If it makes you happy put somethin
From: Evgeniy Polyakov <[EMAIL PROTECTED]>
Date: Tue, 16 May 2006 10:59:23 +0400
> And what if we use ESP which would place it's hashed sequence number as
> port?
If it makes you happy put something like:
case TCP:
case UDP:
case SCTP:
case DCCP:
...
On Tue, May 16, 2006 at 10:59:23AM +0400, Evgeniy Polyakov ([EMAIL PROTECTED])
wrote:
> On Mon, May 15, 2006 at 11:57:12PM -0700, David S. Miller ([EMAIL PROTECTED])
> wrote:
> > From: Evgeniy Polyakov <[EMAIL PROTECTED]>
> > Date: Tue, 16 May 2006 10:19:09 +0400
> >
> > > +static int netchannel
On Mon, May 15, 2006 at 11:57:12PM -0700, David S. Miller ([EMAIL PROTECTED])
wrote:
> From: Evgeniy Polyakov <[EMAIL PROTECTED]>
> Date: Tue, 16 May 2006 10:19:09 +0400
>
> > +static int netchannel_convert_skb_ipv4(struct sk_buff *skb, struct
> > unetchannel *unc)
> > +{
> ...
> > + switch (
From: Evgeniy Polyakov <[EMAIL PROTECTED]>
Date: Tue, 16 May 2006 10:19:09 +0400
> +static int netchannel_convert_skb_ipv4(struct sk_buff *skb, struct
> unetchannel *unc)
> +{
...
> + switch (unc->proto) {
> + case IPPROTO_TCP:
...
> + case IPPROTO_UDP:
...
Why do
Let me also bring attention to another netchannel implementation.
Some design notes [blog copypasts, sorry if it is out of sync sometimes].
First of all, do not use sockets. Just forget that such interface
exists.
New receiving channel abstraction will be created by special syscall,
which allows
