Re: why kill pppd but pppoe-server die?

[James Cameron <james.cameron@hp.com>]

On Thu, May 22, 2008 at 10:54:30AM +0800, cyx_mail wrote:
> (1) when a pppoe connections established, pppoe-server spawns a child
> process "pppd".

Okay.

> I modified the pppoe-server code a little, and let pppoe-server send
> SIGTERM to the child process.

Okay.  If pppd has not detached, then the SIGTERM will also be delivered
to your pppoe-server process.  You must handle this signal.

> pppoe-server die not the child pppd die.

Your observation matches.

> (2) But if I use command "kill pppd pid" in SHELL to kill the child
> pppd, the result pppd die is what I expected.
> 	why?

I don't know.  This is a question about pppoe-server and your current
source code.  Are SIGINT and SIGTERM handled differently by pppoe-server
at different times?  What version of pppoe-server source are you using?
I've just looked at rp-pppoe-3.8/src/pppoe-server.c and the signals are
both handled by exiting, in termHandler().

-- 
James Cameron                         http://quozl.netrek.org/
HP Open Source, Volunteer             http://opensource.hp.com/
PPTP Client Project, Release Engineer http://pptpclient.sourceforge.net/