Re: kernal panics -> hardware issue?



D P Schreber wrote:
On 2006-07-13, Clever Monkey <clvrmnky.invalid@xxxxxxxxxxxxxxxxxxx> wrote:
This problem is almost certainly a firmware/driver issue. Similar backtrace info is like a rash all over Apple Discussions, and the jury is out on the exact cause, though there is a strong indication this is at least partially hardware related.

One thing I should have mentioned earlier is that during the first month
or so, while I was running 10.4.6, I had one kernel panic. Since
upgrading to 10.4.7, which apparently a installed a new darwin kernel
(8.7.1 vs 8.6.1), I've had three.

Ok. Could be coincidence, or could be a driver/device interaction. Or, since we are seeing problems with mbufs, something was changed in pure kernel code that interfaces with the network. We need some smart Apple engineer to put the external debugger on one of the offending boxes and see what the heck is going on in there!

The only real hint we have is that it crashed at least once in com.apple.iokit.IOPCIFamily, and the kernel looked like it was doing something with the graphics module when it got a page fault.

That one happened while Apple's Flurry screen saver was on. Nothing
much else was going on at the time, though there were some active
network connections (see below).

Ah, that explains that backtrace item, then. I'll assume this is a result of the PCI stuff going pear-shaped when the memory it was sitting on was clobbered by something else (perhaps mbufs).

Run with Ethernet only for awhile and see if the panics continue. There are indications that the AirPort device (at some point) may be at fault. There is some speculation that larger frames or more traffic over WPA 801.11g might be the culprit. Some people have reported success replacing the Airport card with one that has newer firmware, though others have indicated that this is not the case, and they required a logic board exchange.

This is interesting, because most often I use the MPB at home, and when
I run at home, I'm always running wireless, at least up until now. When
I bring it to work I run wired, and no crashes have happened there. But
I don't run that way very often.

At this particular moment I'm at home and I have a log cat5 cable
connected to the nat-router. In general it's somewhat awkward to run
this way. I'll give it a shot though.

I know it, but this seems to be the common point of failure for many. Those multiple refs to mbufs points fingers directly at networking of some sort, and the AirPort hardware is somewhat more variable in terms of kernel access and firmware releases than the ethernet device.

If we tend to get no panics with no traffic going over the wireless device, at least it continues to point in the same general direction as "mbufs going pear-shaped, stomping on memory, raise hardware interrupt, panic!"

Not much you can do to fix it, but at least it is a better data point.

I'd carefully reseat all the devices you can get at, and reinstall to the default OS.

I'm not enough of a hardware geek to fiddle with anything other than
memory, but I'll take a look at the dimms. I suppose I should consider
a reinstall. Kind of a pain, but then again so is sending it to Apple.

The reinstall may be overkill, but it proves to Apple that a generic machine has the same problems. Sort of political, but you may have to build a case for a real repair attempt. Start keeping notes!

Is the Airport card user-accessible on this model? I'd make sure you cleaned the contacts and reseated that, as well. May as well get the easy stuff out of the way. When I was a firmware hacker in another life these sorts of intermittent problems were almost always related to noisy connections. And that was industrial hardened equipment. Keeping contacts clean is a real industry challenge, though a little less so for commercial end-user equipment.

Good luck
.



Relevant Pages

  • Why 10.6 So Soon? A New Kernel? Could it be Xen? Arguments for and against
    ... The speed of the new Mac OS release has me thinking. ... Then I keep coming back to the reasons Apple said they are releasing ... stuff, down to the kernel. ... When Mac OS X was first released, the OS was built around the Mach kernel. ...
    (comp.sys.mac.advocacy)
  • Re: unable to send mail from UNIX
    ... Mac OSX is FreeBSD with an Apple ... >> GUI, for the purposes of any Unix question. ... > For example, apple itself then says it also used the mach kernel, which ...
    (comp.unix.admin)
  • Re: The Number One Macintosh Myth...
    ... nVidia chips Apple uses, as if they are unique to Apple, and aren't ... The fact that Apple extremely limits your choice of Wintel parts ... Mac OS X uses the XNU kernel, ...
    (comp.sys.mac.advocacy)
  • Re: Snow Leopard 64-bit Rumor Mongering!
    ... Sermo Malifer stated in post h6f654$v29$2@xxxxxxxxxxxxxxxxx on 8/18/09 2:23 ... You Apple fans may think 64 bit is unnecessary, ... address more than 4GB of RAM, in spite of the kernel being 32 bit. ... And if you have a chance to educate me about a topic I am wrong about I hope ...
    (comp.sys.mac.advocacy)
  • Re: Upgrades for an ibook 800 G3
    ... Are you sure your Airport card isn't the Extreme version? ... updating your router to 802.11g ... Do you have anything that uses Bluetooth? ... socket, you can get a D-Link USB Bluetooth Adapter, that the Apple ...
    (comp.sys.mac.portables)