For all those interested in my little routing problem (I do have 1.5 kids
which could be the problem) we have somewhat solved it.
The problem comes from a test in the kernel (only tested in 2.0.x) for a
0.0.0.0 ip address to handle address hunting (eg bootp). If there is a
device configured (up) on the system with a 0.0.0.0 ip address then all
addresses will be recognised as local and the local arp address returned.
However, addresses on the ppp link do not get handled in the same way and
appear to be immune from this problem. As diald configures the slip device
to have a 0.0.0.0 address when the ppp link is up all addresses not destined
for the ppp link are matched to the local arp address. To overcome this
problem we commented out the line and the end of route.c which ifconfigs the
slip address to 0.0.0.0. This removed the connectivity problem but left two
default routes in the routing table. To solve this we first removed any
existing 0.0.0.0 route in the route table (using addroute) before adding the
new ppp route. i.e route del 0.0.0.0. There is still another route for the
dormant slip connection in the routing table but as this uses a reserved
provate address it is unlikely to be a problem.
This may not be the most elegant solution but it solves the problem.
Dave
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]]
Sent: 04 May 1999 16:03
To: [EMAIL PROTECTED]
Subject: Routing problem
Hi,
I have just installed diald and have some problems.
I hope this is the correct email address for the mail list as the "Welcome
to linux-diald" does not actually tell you what the address is. (After 1
hour a help request still hasn't come back). Also, all links to the mailing
lists archives lead to nowhere so I hope this problem isn't common.
Anyway,
I have successfully setup a pppd connection to an ISP using a dynamic local
and remote IP address. This connection is used to connect a local LAN (using
part of the reserved private address space) to the Internet. The connecting
machine also runs DNS for the local LAN and IP masquerading is used to
connect the local machines to the Internet. Having successfully set this up
I decided to used diald for on demand internet connections as local calls
are timed.
I have downloaded diald0.99, which I assume is the latest, and have set the
machine up accroding to the information I could find. I use addroute, ip-up
and ip-down to handle the routing problems associated with using IP
masquerading. Using diald I can successfully connect to the internet on
demand from the access machine but lose all contact to the local LAN
machines once the connection is setup. An inmvestigation of the routing
tables showed the same configuration for both my manual pppd connection and
diald. The only difference I could find was an error message in
/car/adm/debug
machine kernel: ARP: arp called for own IP address
Further inspection using TCPDUMP showed that a ping from the access machine
to another local machine was being sent but not reply was received. An
inspection of the ping packet showed that the access machine was using it's
own MAC address as the destination address which I assume is why the error
was generated in the log file. The problem goes away as soon as diald
terminates the connection. I am not using proxyarp and turning it on makes
no difference. The ping packet has the correct IP addresses, only the
destination MAC address is in error.
Does nayone know why this is happening and how I fix it.
The machine is running Slackware 3.5 (2.0.34)
Cheers
Dave Atkinson
-
To unsubscribe from this list: send the line "unsubscribe linux-diald" in
the body of a message to [EMAIL PROTECTED]
-
To unsubscribe from this list: send the line "unsubscribe linux-diald" in
the body of a message to [EMAIL PROTECTED]
