bt on innd problem in ov3.c:242

Joe St Sauver JOE at OREGON.UOREGON.EDU
Wed Mar 8 15:52:40 UTC 2000


>Date: Wed, 08 Mar 2000 19:08:11 +0900 (JST)
>From: Katsuhiro Kondou <kondou at nec.co.jp>
>Subject: Re: bt on innd problem in ov3.c:242
>To: inn-workers at isc.org
>
>} A backtrace of the stack at the time innd was crashing on me had been 
>} requested; here's what gdb says...
>:
>} #4  0xdf6d2190 in syslog () from /usr/lib/libc.so.1
>} #5  0x5981c in GROUPfind (group=0x0) at ov3.c:242
>} #6  0x59990 in tradindexed_groupadd (
>}     group=0x700008 "3com.totalservice.support", lo=1455, hi=1826, 
>}     flag=0xdf1c0030 "y\n3com.totalservice.support.courier 0000000885 0000000846
>} y\n3com.totalservice.support.netserver 0000001228 0000001189
>} y\n3com.totalservice.support.radius 0000000594 0000000570
>} y\n3com.totalservice.suppo"...) 
>}     at ov3.c:277
>
>I cannot find syslog() at line 242 in ov3.c.
>I suspect the code you were running was different
>from CURRENT one.

This was from the inn-2000-02-28_03-02 snapshot...

And the chunk of code looked like:

    grouphash = Hash(group, strlen(group));
    memcpy(&i, &grouphash, sizeof(i));

    loc = GROUPheader->hash[i % GROUPHEADERHASHSIZE];
    GROUPremapifneeded(loc);

    while (!GROUPLOCempty(loc)) {
        if (GROUPentries[loc.recno].deleted == 0) {
            if (memcmp(&grouphash, &GROUPentries[loc.recno].hash, sizeof(HASH))
== 0) {
                return loc;
            }
        }
        loc = GROUPentries[loc.recno].next;
    }
    return GROUPemptyloc;


If I am to believe my editor, the "if" that appears immediately after the 
"while" statement is line 242...

Shrug. I had to get the box back up, so I recompiled w/o large file support,
blasted the history, and rolled on.

Thanks for at least looking at it for me.

Regards,

Joe



More information about the inn-workers mailing list