Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 8 Apr 2001 13:01:16 +0200 (MEST)
From:      Martin Husemann <martin@duskware.de>
To:        martin@night-porter.duskware.de
Cc:        Brian Somers <brian@Awfulhak.org>, Martin Husemann <martin@duskware.de>, freebsd-isdn@FreeBSD.org, joerg@FreeBSD.org
Subject:   Re: PPP encapsulation
Message-ID:  <200104081101.f38B1GT00379@night-porter.duskware.de>
In-Reply-To:  "from (env: martin) at Apr 7, 2001 02:25:16 pm"

next in thread | raw e-mail | index | archive | help
Ok, tried it: did not work.

Actually the framing implemented by if_spppsubr.c is not a complete
implementation of RFC 1662 framing, it does not add the flag byte (0x7e) nor
FCS. (The FCS was my main concern.)

So it only prepends one byte adress (0xff for PPP = all stations), and a
controll byte (unnumbered HDLC information element, 0x03).

Not doing this makes the PPP conversation not work. Anyone know where this
"sync PPP header" (see struct ppp_header in if_spppsubr.c) is specified?

While there I noted that some comments and the statisitcs (bytes output, bytes
received) always add 3 bytes, from comments indicating exactly the missing
parts of RFC 1662 framing (one byte "flag", two byte FCS).

IMHO both the comments and the +3 calculation are wrong.

Maybe if in CISCO encapsulation mode, the sync hardware does add these, which
means we should not add 3 bytes when in PPP mode. Probably best solution
would be to add an element in strupt sppp giving the constant byte number
added by hardware and always using that (which would be 0 in the PPP over
ISDN case, 3 in the sync hardware case and yet something else in the
application I'm currently working on).

Anyone care to sanity check this?


Martin

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-isdn" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200104081101.f38B1GT00379>