I committed the following:
#v+
[gnumach]/
2006-04-27 Richard Braun <[EMAIL PROTECTED]>
Manuel Menal <[EMAIL PROTECTED]>
* device/if_hdr.h (struct ifnet): Added new members `if_snd_port_list'
and `if_snd_port_list_lock'.
* device/net_io.c: Reworked to improve
On Wed, Apr 19, 2006 at 04:32:57PM +0200, Marco Gerards wrote:
> Richard Braun <[EMAIL PROTECTED]> writes:
> > Here is the NETF filter that would be used in pfinet :
> > static short ether_filter[] =
> > {
> > NETF_IN, /* header */
> > NETF_PUSHLIT | NETF_NOP,
> > 1
> > };
>
> Is it possible
Richard Braun <[EMAIL PROTECTED]> writes:
> NETF filters are considered to be native filters, so there is no macro
> for them (a requirement for all non-native filters is to be non-zero).
>
> Macros were also added in include/device/bpf.h :
> #define BPF_BEGIN NETF_BPF
> #define BPF_IN
OK, so, after getting some feedback, here is a new patch (the change in
kern/queue.h was inappropriate, and a useful comment was added).
I'm also going to describe the changes brought to the interface to packet
filters more clearly.
The main changes to the interface are visible in include/device/n
Hello,
Well, again, with the patch this time ;-).
Here is a patch that extends the previous one (it replaces
18_net_filter_fix.patch). The main improvement is the creation of
a new port list for egress packets. As said in my previous message, we
had problems doing this before, but they are now so
Hello,
Here is a patch that extends the previous one (it replaces
18_net_filter_fix.patch). The main improvement is the creation of
a new port list for egress packets. As said in my previous message, we
had problems doing this before, but they are now solved.
I had to modify kern/queue.h, and thi
implied that we were not able to solve (like rearranging the
filters queues to separate ingress and egress filters, adapting the GC
processing, etc...). However, the GNU Mach interface to packet filters
should be modified to allow listeners to specify the direction of the
packets they want to
