Re: possible bug

From: Thomas Osterried (gbhbifp.lmip@dircon.net)
Date: Thu May 10 2001 - 02:31:27 EEST

  • Next message: Kjell Jarl: "uscc endelss loop"

    > > I think it is failing on this:
    > > skb2->nh.raw < skb2->data || skb2->nh.raw >= skb2->tail
    [..]
    > It was awhile back, but I thought this was related somehow to the fact that
    > with ax25 the tcp/ip data doesn't start at a constant offset. I think, because
    > of the digipeat stuff. I tried tracking down the 'is buggy' message about a
    > year ago, and I think I convinced myself that it was harmless at the time. Though
    > my memory is vague on this.

    i also tracked this down, when i was experimenting with wampes while
    attaching to the soundmodem sm0-device (something, that worked under
    kernel 2.0.x).
    again, i saw it, while debuging krnlif.c (attaching to the kernel ip from
    wampes, that stopped working when 2.2.x kernels came up (SOCK_PACKET problem)).
    and last but not least, while testing the bpq0 features (ax25 over ethernet
    mac layer), i saw these messages.

    i must admit, that i never understood, what's *really* going on,
    but i found exactly that piece of code.
    well, the author sais "Code, following below is wrong. [...] so that the
    second statement is just protection against buggy protocols.", and i
    believed it was a hack to make things work, and may be rewritten some time..

    hmm..

    73,
            - thomas dl9sau
    -
    To unsubscribe from this list: send the line "unsubscribe linux-hams" in
    the body of a message to nmm@uni-sb.de



    This archive was generated by hypermail 2b30 : Thu May 10 2001 - 02:37:17 EEST