lwip-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [lwip-users] PPP negotiation frame length and sio_read() inco mpatib


From: Wilson, Clive
Subject: RE: [lwip-users] PPP negotiation frame length and sio_read() inco mpatibility
Date: Wed, 17 May 2006 18:00:52 +0100

Guille,

Thanks for your help. Does your target application require the PPP link to be able to be brought up and down on demand? Do you find there is a problem with bringing the stack down, then being able to start it all up again? The reason I ask is that I have seen others in lwip-users mentioning problems in this regard. If you are using a modem in your system, how do you signal to PPP that you have detected a loss of connection (other end hung up their modem)?

Bye for now,

CTW

> -----Original Message-----
> From: Guillermo Prandi [mailto:address@hidden]
> Sent: 17 May 2006 15:01
> To: 'Mailing list for lwIP users'
> Subject: RE: [lwip-users] PPP negotiation frame length and
> sio_read() inco mpatibility
>
> Wilson, you should block for at least one character in order
> to avoid excesive CPU usage (which will of course slow down
> everything).

> I only had to modify ppp.c:sifup() in order to make it work
> up to the UDP level (I haven't tested TCP). Read my other
> mail regarding bug #16602 for details. Otherwise,
> implementation was very straightforward.

> Guille

> -----Original Message-----
> From: Wilson, Clive [mailto:address@hidden]
> Sent: Wednesday, May 17, 2006 10:50
> To: 'Mailing list for lwIP users'
> Subject: RE: [lwip-users] PPP negotiation frame length and
> sio_read() inco mpatibility
>
>
>
> Guille,
>
> Thanks for your reply. Actually, this morning I took a fresh
> look at the PPP code and realised that the only way the
> sio_read() call in pppMain() could possibly work was if it
> was non blocking. So I changed my implementation of
> sio_read() so that it returns immediately, irrespective of
> how much data it has retrieved. The good news is that an LCP
> dialogue of sorts is happening - however, the peers get as
> far as agreeing the Auth type, then the process repeats with
> the original ID number again. Ultimately one end gives up and
> the process fails. I also find that throughput is very slow -
> this may be the reason for the above problem?
>
> Did you find you had to change any PPP code (LCP, Auth, IPCP
> etc) in order to get the whole thing working? I'm not at all
> confident that the PPP code as shipped will work OK - I
> figure that very few people are actively using it. Either
> that or their not prepared to merge their findings into the
> current LwIP development 8-)
>
> Bye for now,
>
> CTW
>
> > -----Original Message-----
> > From: Guillermo Prandi [mailto:address@hidden]
> > Sent: 17 May 2006 14:35
> > To: 'address@hidden'
> > Subject: RE: [lwip-users] PPP negotiation frame length and
> > sio_read() inco mpatibility
> >
> > Hi, Wilson. I got it working in my Windows test
> implementation by not
> > returning unless at least one byte was read. >From the
> source code I
> > interpreted that sio_read should return AT MOST the specified
> > number of
> > characters. Additionally, Windows lets you specify an
> inter-character
> > timeout for the read operation, so you can keep reading if
> > more characters
> > keep coming and thus reduce the processing overhead. You
> > could implement
> > something alike (some UARTs implement such timeout too).
> >
> > Guille
> >
> >
> > _______________________________________________
> > lwip-users mailing list
> > address@hidden
> > http://lists.nongnu.org/mailman/listinfo/lwip-users
> >
>
>


reply via email to

[Prev in Thread] Current Thread [Next in Thread]