Hi, all,
For those who have ppp problem, here is my test report. There are a
good news and a bad news. The good one is that my PCMCIA modem card works
and is very stable now. The bad news is I can not determine exactly where
this bugfix comes from.
First I briefly describe my environment. I am currently a graduated
student of National Taiwan University. When I stay at school, I always use
PCMCIA Ethernet card to access the Internet. When the winter vacation
came, I went back to my home town for two weeks. So I bought a PCMCIA
modem card such that I can use dial up during the vacation.
At my home town, I can use the new modem card, but the connection
worked for just five minutes and then suddenly hanged up. The weird thing
I noticed is the content in ppp.log:
Feb 10 22:14:28 newton pppd[449]: sent [LCP EchoReq id=0x8 magic=0xa88cda89]
Feb 10 22:14:28 newton pppd[449]: rcvd [LCP EchoRep id=0x8 magic=0x0]
Feb 10 22:14:58 newton pppd[449]: sent [LCP EchoReq id=0x9 magic=0xa88cda89]
Feb 10 22:14:58 newton pppd[449]: rcvd [LCP EchoRep id=0x9 magic=0x0]
Feb 10 22:15:28 newton pppd[449]: sent [LCP EchoReq id=0xa magic=0xa88cda89]
Feb 10 22:15:58 newton pppd[449]: sent [LCP EchoReq id=0xb magic=0xa88cda89]
Feb 10 22:16:28 newton pppd[449]: sent [LCP EchoReq id=0xc magic=0xa88cda89]
That is, the pppd can not receive acknowledge of EchoReq after some time.
After I was back to school, I reported my situation on the list and
browsed the bug tracking system on http://www.debian.org. Unfortunately,
it seems this problem is pretty involved. Ppp maintainer thought this is
actually setserial problem, but setserial maintainer guessed it is
pcmcia-cs that caused this problem. After a while, the setserial
maintainer reported the PCMCIA modem problem has been fixed. So, I decided
to try those upgraded packages (ppp, setserial) yesterday and found the
modem card works fine.
Hence, there are two possibilities:
1. The ISP circuits and equipments at my school is better than those at my
home town. So this modem card can now work at my home town.
2. The bug is fixed by the upgraded setserial package. However, since I
usually use Ethernet access, I don't know which revision fixed this
problem.
I did not use the remedies mentioned by George Bonser yesterday. But since
the problem is already solved, so I can not determine whether his
solutions can also fix the problem.
So that's my report. Any idea?
Jonathan Chang
