[2197] in linux-net channel archive

home help back first fref pref prev next nref lref last post

Re: LCP echo problems (again)

daemon@ATHENA.MIT.EDU (linuxsys@ssg.com)
Thu Mar 21 10:54:10 1996

Date: 	Thu, 21 Mar 1996 14:10:48 GMT
From: linuxsys@ssg.com
To: Paul Mackerras <paulus@cs.anu.edu.au>
Cc: longyear@netcom.com, linuxsys@ssg.com, linux-net@vger.rutgers.edu
In-Reply-To: <199603202247.JAA11460@sirius.anu.edu.au>

Paul Mackerras writes:
 > > Linux pppd version 2.2 sends an empty CCP frame if you don't have any
 > > compressors defined.
 > 
 > If no compressors are enabled, pppd is supposed to bring up CCP in
 > "silent" mode, so it won't send a CCP frame unless it sees one from
 > the peer.  It needs to be able to send an empty CCP frame in the case
 > that the peer wants us to compress stuff going to it, and we're
 > willing, but we don't want stuff coming to us compressed.
 > 
 > It looks to me like the problem is that pppd is not checking with the
 > kernel which compression options are supported before making the
 > decision about whether to open CCP in silent mode or not.  This is a
 > bug which can be fixed by changing the line in ccp_open which says
 > 
 > 	if (!ANY_COMPRESS(ccp_wantoptions[unit]))
 > 
 > to
 > 	ccp_resetci(f);
 > 	if (!ANY_COMPRESS(ccp_gotoptions[unit]))
 > 
 > Also, I think we need a -ccp option (or something) to disable CCP
 > completely.
 > 

Al, Paul, which route should I take when trying to stay connected with
my ISP for more than ~5 minutes at a shot?  Delete the CCP code or try
this patch?  I agree that -/+ccp is a very good idea.

Rick



home help back first fref pref prev next nref lref last post