My system (RH6.0, kernel 2.2.5-15, configured with SLIP in the kernel
and tap as a module) shows some very strange behaviour when I try to use
the tap interface for diald's proxy. I checked the FAQ and deja.com with
no luck�
I have a working connect script (a stripped-down version of the one in
/usr/lib/diald), such that running "/usr/sbin/pppd /dev/ttyS0 155200
file /etc/ppp/options-nodiald connect /etc/ppp/scripts/connect" runs
like charm. The file /etc/ppp/options-nodiald is a pppd options file
that includes such things as lock, defaultroute, etc. Another file
/etc/ppp/options contains just the options needed by pppd when running
under diald control, and diald.conf has the name of this very same
connect script as its "connect" parameter.
Now for the weird part: if I don't create any tap devices (i.e. I don't
include any tap<n> aliases for ethertap in conf.modules) diald defaults
to the slip proxy, and works OK (except for CCP complainig about a
"timeoutsending Config-Requests" after connecting, but I guess I'll
figure that one out). But if I DO create at least one tap device, diald
uses that instead of slip (OK so far), but then somehow manages to break
chat: the same chat invokation that works when using the slip proxy dies
with the message "Can't get terminal parameters: Invalid argument" when
run using tap. The same problem occurs if I compile tap into the kernel.
This is not really a problem for me, at least not yet as I don't use any
other software that requires tap devices, but it really intrigues me.
Has anybody else seen this problem before?
Federico Heinz
-
To unsubscribe from this list: send the line "unsubscribe linux-diald" in
the body of a message to [EMAIL PROTECTED]
