Re: child process exited due to signal 6

From: Brian Beuning <bbeuning@dont-contact.us>
Date: Tue, 26 Oct 1999 23:26:38 -0400

In statHistBin(), I changed this function to look like this

    v -= H->min; /* offset */
    if (v <= 0.0) /* too small */
        return 0;
    d1 = H->val_in(v);
    d2 = H->scale * d1 + 0.5;
    d3 = floor(d2);
    debug(62, 1) ("statHistBin: d1=%f d2=%f d3=%f\n", d1, d2, d3 );
    bin = (int) floor(H->scale * H->val_in(v) + 0.5);

    debug(62, 1) ("statHistBin: bin=%d\n", bin );

When I run squid like this, the output says:

1999/10/26 23:21:53| statHistBin: d1=35.000000 d2=35.500000 d3=0.000000

So how come floor(35.5) == 0.0 ??

When I write a little test program...
on alpha (RH 6.0) floor(35.5) is 0.0
on a 486 (RH 4.?) floor(35.5) is 35.0

Hmmm... Sounds like an alpha bug.

This is RH 6.0 on an alpha after installing a new kernel.

Brian

Brian Beuning wrote:

> After turning on "debug_level ALL,1" this in the cache.log file.
>
> assertion failed: StatHist.c:91: "statHistBin(H, max) == H->capacity -
> 1"
>
> I am recompiling to figure out the value of H->capacity.
>
> Brian
>
> Brian Beuning wrote:
>
> > I had been running squid (2.2.STABLE1) fine for a couple
> > of months with RH6.0 and linux 2.2.5-22 on an alpha.
> >
> > Now I built kernel 2.2.12 (without the tulip.c driver, to
> > fix my network wedgie problem) and now squid will not
> > start.
> >
> > The squid log files contain no new entries. The system
> > /var/log/messages
> > file has lines like
> >
> > Squid Parent: child process 13232 started
> > Squid Parent: child process 13232 exited due to signal 6
> >
> > My best guess is I did not configure something into the kernel that
> > squid depends on. Any suggestions what this might be?
> >
> > Thanks,
> > Brian Beuning
Received on Tue Oct 26 1999 - 21:34:51 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:49:05 MST