Re: More pre-2.0.31-1 oops's....

Doug Ledford (dledford@dialnet.net)
Tue, 17 Jun 1997 17:32:38 -0500


--------
> On Tue, 17 Jun 1997, Doug Ledford wrote:
>
> > --------
> > > Again, I got a pair of these just a few minutes apart. Since the last
> > > time, I've upgraded in.identd to 2.5.1, but it would seem it must be a
> > > kernel bug, since a process running as nobody shouldn't be able to screw
> > > up the kernel. Fortunately, the system is still running with no apparent
> > > ill effects.
> >
> > Try a "cat /proc/net/dev" and then tell me no ill side effects :) No, at
> > this point, your internal kernel device list (or the proc representation of
> > said list) is nice and toasty. Fine if you don't need to touch the devices.
>
> Toasted in what way? This is a simple (though large box) with only lo and
> eth0.
>
> yoda:~# cat /proc/net/dev
> Inter-| Receive | Transmit
> face |packets errs drop fifo frame|packets errs drop fifo colls carrier
> lo:7134930 0 0 0 0 7134930 0 0 0 0 0
> eth0:136510670 0 0 0 0 115845034 0 0 0 643205 0

You got lucky then. When I saw any get_netinfo oopses, it always meant my
/proc/net/dev file was toasted, but then again, that machine had roughly 116
entries. Anyway, this particular problem has always meant (on my machines
at least) that there was something wrong with the internal device table. In
this case, it may not be dying on a cat of /proc/net/dev, but it is likely
that some structure a little bit deeper down for one of the two interfaces
has a bad pointer and is resulting in the oopses you are getting. One thing
you can do is try to cat each of the files in /proc/net and see if any of
them result in oopses. That would probably let you know where the bad
pointer is at least.

> yoda:~# ifconfig
> lo Link encap:Local Loopback
> inet addr:127.0.0.1 Bcast:127.255.255.255 Mask:255.0.0.0
> UP BROADCAST LOOPBACK RUNNING MTU:3584 Metric:1
> RX packets:7134964 errors:0 dropped:0 overruns:0
> TX packets:7134964 errors:0 dropped:0 overruns:0
>
> eth0 Link encap:10Mbps Ethernet HWaddr 00:00:C0:E8:D9:9E
> inet addr:205.229.48.17 Bcast:205.229.48.255 Mask:255.255.255.0
> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
> RX packets:136511062 errors:0 dropped:0 overruns:0
> TX packets:115845361 errors:0 dropped:0 overruns:0
> Interrupt:10 Base address:0x310 Memory:d4000-d8000
>
> If there's a problem here, it's eluding me...but I'm ill and quite tired.
>
> > release of 2.0.30. Dave's pre-2.0.30-2 patch does not have this problem.
> > Somebody should probably do a diff between Dave's last pre-2.0.30 tree and
> > Linus' official 2.0.30 tree to see where things might have gone wrong.
>
> Actually, I'd like to see a pre-pre-patch-2.0.31-3 that includes
> pre-2.0.31-2 plus the memory management fixes...so I can upgrade yoda.
>
> This mess with 2.0.30-x has me wondering whether its time to start looking
> at BSD again.

You know....I had a system that was crashing about once every day. It would
lock up and someone would have to reset it. I thought I might try FreeBSD
to see what would happen. Well, it didn't lock up and require resets any
more, instead, it was rebooting itself about once every two hours due to a
kernel problem. So, I ended up putting linux back on that box, dumping a
certain driver, and now it hasn't given me another problem. I'm not
convinced that FreeBSD is the answer to what ails you :) But, that's a bit
off-topic here so I won't say any more :)

-- 
*****************************************************************************
* Doug Ledford                      *   Unix, Novell, Dos, Windows 3.x,     *
* dledford@dialnet.net    873-DIAL  *     WfW, Windows 95 & NT Technician   *
*   PPP access $14.95/month         *****************************************
*   Springfield, MO and surrounding * Usenet news, e-mail and shell account.*
*   communities.  Sign-up online at * Web page creation and hosting, other  *
*   873-9000 V.34                   * services available, call for info.    *
*****************************************************************************