Guillaume Rousse wrote:
I have good guesses but not enough knowledge and time to accurately compare the source codes algorythm of both versions.Le Dimanche 24 Novembre 2002 14:22, Lea Gris a écrit : [..]Why do you think it is a bug, and not a normal behavior from pppd ? After all, this is just standard pppd 2.4.1 with pppoatm patch.About the pppd 2.4.1 from plf it has a annoying bug avoiding it to be able to keep the ATM device open and retry forever to reconnect if the ADSL link drop. It is the case every 24h in France.
BTW, do you know the reason of this drop ?
So :
When France Télécom drops the ADSL line every 24h pppd 2.4.1 just exit with signal 11 besides the fact it has the persist option.
I'v done several test by manuyally unplugging the ADSL line and it is like if the ADSL link stay off for more than 30s (estimated) the pppd exit with signal 11. If the ADSL line drop for a shorter period it is able to restart a new ppp session (besides the modem realy dropped its synchronization from the DSLAM during that time )
with the persist option pppd should keep on tryining to communicate over the device even if the device is in an offline like state.
I can't tell you what changed between pppd 2.4.0-2 and pppd 2.4.1 that makes the later exit signal 11 when the ADSL link goes down for more than 30s and therefore the ATM device may return some kind of unavailable status or offline status the pppoatm.o plugin or pppd 2.4.1 can't manage this status.
pppd 2.4.0-2 just keep trying to get tha ATM connection forever whent it has the persist option. This gives enough times for the ADSL link to goes up and the modem to synchronize again.
While on a serial line driven modem. The serial bus remain connected even if the phone line drop and the modem hang on the phone line. These are different behaviors.
--
Léa Gris
() Campagne du ruban texte brute contre les courriels en HTML,
/\ contre les pièces jointes Microsoft.
