When specifying options "-u -o XXX" to pppoe-server internal PPPoE session offsets really set off, but pppX does not. Problem is, there are no specific handling for passing of "unit" option to pppd, lack of which leads to mixing pppX from different PPPoE servers and clients. Attached patch propagates the session offsets to "unit" for pppd, thus solving the problem. Reproducible: Always Steps to Reproduce: 1. Start pppoe-server: pppoe-server -u -o 10 -k 2. Track the syslog for pppoe-server messages 3. Connect the client Actual Results: Something like: pppoe-server[26393]: Session 11 created for client [mac] ([ip]) on eth0 using Service-Name 'service' ... pppd[26393]: pppd 2.4.4 started by root, uid 0 pppd[26393]: using channel 27 pppd[26393]: Using interface ppp1 pppd[26393]: Connect: ppp1 <--> eth0 Expected Results: Correct behavior (observed after the patch): pppoe-server[26652]: Session 11 created for client [mac] ([ip]) on eth0 using Service-Name 'service' ... pppd[26678]: pppd 2.4.4 started by root, uid 0 pppd[26678]: using channel 27 pppd[26678]: Using interface ppp10 pppd[26678]: Connect: ppp10 <--> eth0 Offsets are either for pppX or for session.
Created attachment 140228 [details, diff] Proposed patch Proposed patch which solves the problem
Don't use this. *** This bug has been marked as a duplicate of bug 204451 ***
Jakub, AFAIK, rp-pppoe is the only solution that provides PPPoE server support and it works! I can't see the point of rejecting the bugfix on basis "don't use the package" without specifying what else should user do. From my POV closing this bug and marking it as duplicate of "removal request" is inadequate. Please reopen it.
Reopening as of marking "duplicate bug" as invalid.
Your patch wasn't complete. startPPPDUserMode() also needed this modification. Fixed in rp-pppoe-3.8-r2. Thanks!
Verified, thanks.