Hallo,
On Mon, Jan 17, 2000 at 04:24:57PM +0100, [EMAIL PROTECTED] wrote:
> Hello again :)
> On Sun, Jan 16, 2000 at 02:37:09PM +0100, Gigi Sullivan wrote:
> > What I said just looking (ok too fast ;)) at the new stuff (2.2.X)
> > is that you can access and get info about the route entries, but
> > one at a time. Again there's no simple (like a single msg/ioctl) to
> > get all the list.
> > However for this part (2.2.X stuff) I could be really wrong, so
> > forgive me :)
>
Ok, thx a lot.
I looked in the source and I found something I guess
I don't understand right well.
Into /usr/src/linux/net/ipv4/af_inet.c (about line 883) you can find:
case SIOCADDRT:
case SIOCDELRT:
case SIOCRTMSG:
return(ip_rt_ioctl(cmd,(void *) arg));
This sounds good but when we look at
/usr/src/linux/net/ipv4/fib_frontend.c (where the ip_rt_ioctl
implementation lives) I didn't find the code necessary to
manage the SIOCRTMSG ioctl call.
All ip_rt_ioctl stuff here is also #ifndef CONFIG_IP_NOSIOCRT.
Which is set per default.
So, my questions are:
1) Does the ip_rt_ioctl (and then the
ioctl(SIOC[ADDRT|DELRT|RTMSG)) disappeared ?
I mean, Is all the routeing staff managed by the
(new) rtnetlink code ?
ADDRT/DELRT still works, but it cannot configure all the new routing
features (policy routing)
RTMSG was never supported.
2) Why the SIOCRTMSG isn't handled ? Is that a miss ? Is it
useless and/or obsolete ?
It does not exist.
If I'm right, however, I have no more need to use ioctl() to
add/del route entries ?
Most people still use ioctl (via ifconfig), but it is an outdated
dead interface
What do you think, however, about the ioctl(SIOCTLSTRT) addendum
for the 2.0.38 kernel ? It could be usefull (even if I'm talking about
the old [but still used] kernel series) ?
I'm afraid I don't think it makes much sense to add a new interface
to the dead 2.0 tree when 2.2+ has a different and better interface.
-Andi
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to [EMAIL PROTECTED]
