inn-stable crash

Russ Allbery rra at stanford.edu
Tue Apr 4 23:23:49 UTC 2000


Ari Heikkinen <aheikin at phaser-atm.pspt.fi> writes:

> I'm running "inn-STABLE_2_2-2000-02-19_03-02" on our news server which
> just happened to crash (coredump) a while ago.

> I did a quick stack trace with dbx for it and here's the result:

> -----------------------------------------------------------------------------
> ttyq1: news at raptor /raid/news/spool/articles > dbx
> /home/news/stuff/inn-STABLE_2_2-2000-02-19_03-02/innd/innd core
> dbx version 7.2.1.3m Dec 23 1998 01:09:37
> Core from signal SIGSEGV: Segmentation violation
> (dbx) where
>>  0 RCreader(0x0, 0x0, 0x10, 0x10067a60, 0x7, 0x0, 0x0, 0x210)
>   ["/home/news/stuff/inn-STABLE_2_2-2000-02-19_03-02/innd/rc.c":495,
>    0x10028afc]

Hm.  That's this block of code:

    /* If not a server, and not allowing anyone, hand him off unless
       not spawning nnrpd in which case we return an error. */
    if ((i >= 0) && !rp->Skip) {
        if ((new = NCcreate(fd, rp->Password[0] != '\0', FALSE)) == NULL)
            return;
        new->Streaming = rp->Streaming;
        new->Skip = rp->Skip;
        new->NoResendId = rp->NoResendId;
        new->MaxCnx = rp->MaxCnx;

The return; line to be precise.  That makes me think that rp->Password was
NULL.  I'm not sure how that could happen, though; it appears to be
properly initialized everywhere to the empty string unless a password is
specified.  I'd need to get some more debugging information about what rp
is set to and the like before being able to figure out what's going on
here.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>



More information about the inn-bugs mailing list