Endian Bugtracker
Endian Issue Tracker

Viewing Issue Simple Details Jump to Notes ] View Advanced ] Issue History ] Print ]
ID Category Severity Reproducibility Date Submitted Last Update
0001463 [Endian Firewall] Network related (VPN, uplinks) major always 2008-11-14 18:15 2009-10-27 13:01
Reporter peter View Status public  
Assigned To peter
Priority normal Resolution fixed  
Status closed   Product Version 2.2-rc3
Summary 0001463: pppcall does not recognize when pppd dialing in using rp-pppoe receives a timeout and exits
Description now PPPoE is the next, which behaves like ISDN and UMTS during dialup.

A system with a PPPoE connection which has no ethernet cable attached, let's the pppd timeout waiting for a PADO. After the timeout pppd exits, but pppcall continues to wait:

0xb7d40bae in __open_nocancel () from /lib/tls/i686/libc.so.6
(gdb) where
#0 0xb7d40bae in __open_nocancel () from /lib/tls/i686/libc.so.6
0000001 0xb7ceb044 in _IO_new_file_fopen (fp=0x8e90908, filename=0x8ea5200 "/var/efw//uplinks//main//main.fifo", mode=Variable "mode" is not available.
) at fileops.c:233
0000002 0xb7ce33d2 in __fopen_internal (filename=0x8ea5200 "/var/efw//uplinks//main//main.fifo", mode=0xb7c5eed4 "r", is32=0) at iofopen.c:93
0000003 0xb7ce5471 in _IO_fopen64 (filename=0x8ea5200 "/var/efw//uplinks//main//main.fifo", mode=0xb7c5eed4 "r") at iofopen64.c:39


uplinksdaemon then kills away the dialup process after 5 minutes, but that's way to long to wait for a pppoE connection which retries 3 times having a timeout of 5 minutes every time.

This causes the backup uplink to come up after more than 15 minutes and that's not acceptable
Additional Information
Tags needsfix
Attached Files

- Relationships

-  Notes
(0001790)
peter (administrator)
2008-11-14 19:31

the problem is within pppd. pppd exits with exitcode 0 when rp-pppoe gets a timeout waiting for PADO
It should exit with an errorcode != 0, I think it should be 10

" 10 The PPP negotiation failed, that is, it didn’t reach the point
              where at least one network protocol (e.g. IP) was running.
"
(0001791)
peter (administrator)
2008-11-14 20:09

status == EXIT_OK witin start_link() before calling connect()
if connect() fails, status is still EXIT_OK, because nobody (the plugins) sets a failure.
I think status must always be != EXIT_OK on start_link()
(0001796)
peter (administrator)
2008-11-17 20:34

moved status = EXIT_NEGOTIATION_FAILED before something happens in start_link()
So if first steps go wrong, it exits with failure

- Issue History
Date Modified Username Field Change
2008-11-14 18:15 peter New Issue
2008-11-14 18:15 peter Assigned To => peter
2008-11-14 18:15 peter Tag Attached: needsfix
2008-11-14 19:31 peter Note Added: 0001790
2008-11-14 20:09 peter Note Added: 0001791
2008-11-17 20:34 peter Note Added: 0001796
2008-11-17 20:34 peter Status new => resolved
2008-11-17 20:34 peter Fixed in Version => 2.3
2008-11-17 20:34 peter Resolution open => fixed
2009-10-27 13:01 peter Status resolved => closed

Copyright © 2005-2008 Endian, SRL. All rights reserved.


Mantis [^]
Copyright © 2000 - 2008 Mantis Group
Powered by Mantis Bugtracker