[pkeene]

After 18 months of relatively trouble free computing I have suffered 3 kernel panics on my iMac, one on each of the last 3 days. The panic logs, example below, seem to point to the internal USB modem as the culprit. Possibly the driver has been corrupted. My question of this learned body is: can I copy the driver off my PowerBook to replace the sick one on my iMac? Both are running 10.3.5. If so, where do I find the driver? I've done a hard drive search on the iMac and haven't found it yet.
It also appears from the panic log that 10.3.5 installed a new version of the kernel which, given the number of posts on the Apple boards about kernel panics since 10.3.5, is possibly related to these problems. However, so far the PB is chugging along happily on the update so maybe that's unrelated.
Your suggestions if you please maestros.
>Thu Aug 19 14:26:22 2004


Unresolved kernel trap(cpu 0): 0x300 - Data access DAR=0x0000000000000004 PC=0x00000000002473D4
Latest crash info for cpu 0:
Exception state (sv=0x19BE9280)
PC=0x002473D4; MSR=0x00009030; DAR=0x00000004; DSISR=0x40000000; LR=0x002473D4; R1=0x079CB710; XCP=0x0000000C (0x300 - Data access)
Backtrace:
0x008CD704 0x00891748 0x008C7A68 0x0088FD40 0x008BBF58 0x008BCE88 0x0087F518 0x008990D0
0x00897E08 0x009A7E24 0x003C90E8 0x003C81EC 0x000D6D44 0x000CDA60 0x000C61C8 0x002452B4
0x00094200 0x00000000
Kernel loadable modules in backtrace (with dependencies):
com.apple.driver.InternalUSBModem(2.6.3)@0x86a000
dependency: com.apple.iokit.IOUSBFamily(2.1.1)@0x3e9000
dependency: com.apple.iokit.IOSerialFamily(6.0.1d20)@0x3c5000
com.apple.driver.InternalModemSupport(2.2.3)@0x9a6000
dependency: com.apple.iokit.IOUSBFamily(2.1.1)@0x3e9000
dependency: com.apple.iokit.IOSerialFamily(6.0.1d20)@0x3c5000
com.apple.iokit.IOSerialFamily(6.0.1d20)@0x3c5000
Proceeding back via exception chain:
Exception state (sv=0x19BE9280)
previously dumped as "Latest" state. skipping...
Exception state (sv=0x1949BA00)
PC=0x9000E88C; MSR=0x0000F930; DAR=0xE00FD000; DSISR=0x42000000; LR=0x0001FD40; R1=0xBFFFF500; XCP=0x00000030 (0xC00 - System call)
Kernel version:
Darwin Kernel Version 7.5.0:
Thu Aug 5 19:26:16 PDT 2004; root:xnu/xnu-517.7.21.obj~3/RELEASE_PPC

panic(cpu 0): 0x300 - Data access
Latest stack backtrace for cpu 0:
Backtrace:
0x000836E4 0x00083BC8 0x0001EDA4 0x00090C60 0x0009406C
Proceeding back via exception chain:
Exception state (sv=0x19BE9280)
PC=0x002473D4; MSR=0x00009030; DAR=0x00000004; DSISR=0x40000000; LR=0x002473D4; R1=0x079CB710; XCP=0x0000000C (0x300 - Data access)
Backtrace:
0x008CD704 0x00891748 0x008C7A68 0x0088FD40 0x008BBF58 0x008BCE88 0x0087F518 0x008990D0
0x00897E08 0x009A7E24 0x003C90E8 0x003C81EC 0x000D6D44 0x000CDA60 0x000C61C8 0x002452B4
0x00094200 0x00000000
Kernel loadable modules in backtrace (with dependencies):
com.apple.driver.InternalUSBModem(2.6.3)@0x86a000
dependency: com.apple.iokit.IOUSBFamily(2.1.1)@0x3e9000
dependency: com.apple.iokit.IOS`
*<
Peter

[Praxis]

Well, I don't understand the technobabble above, BUT kernel panics are almost always caused by hardware issues, most commonly bad RAM (and sometimes loose cards inside the computer).

[berg]

All I know about kernel panics ...
Resolving Kernel Panics
http://www.thexlab.c...rnelpanics.html

Bad Ram & Kernel Panics
http://www.macattorn...or-Kernel-44484
Also, your Ram could be Loose

[DPG4450Guy]

Strange, because in my experience, most kernel panics are attributable to either corrupt kexts or just plain bad programming, or sometimes out-of-date applications.
(Jag to Panther required alot of updates for many applications.)
I've never monkeyed with machines with no-name RAM, so hardware has never played into it.

[pkeene]

I only have the original RAM in the machine, no third party drives and, other than 10.3.5, I haven't installed any software recently. So, yep, a corrupt modem kext is what I think I'm looking at (still haven't found it). Would be useful to have an opinion on whether I could copy the driver from my PB. Thanks for the links but I've been there and done that. The logs seems pretty conclusive about pointing the finger at the modem.
Strangely the panics have only happened once a day, twice while connecting and once when waking from sleep. I've connected about 15 times tonight without problem.
Peter

[Praxis]

The only time I had kernel panics was when my Airport Extreme card inside the PB came loose.

[pkeene]

Don't have one of those either. I'm going to turn in now, it's nearly dawn here. I'll check back tonight.
Peter

[worf]

I have had a series of kernel panics following upgrading to 10.3.5. After each one, I rebooted using safe mode ( holdingdown the shift key) to see if it would boot that way - when it did, I ran Disk Utility and repaired permissions. Most recentl;y, however, I tried something different and booted into single user mode (command-s on startup) and ran fsck about four times, which seems to have cured my machine - a B&W 450 with a Sonnet G4/800 mhz upgrade in it - 768 mb of RAM. I would try it, if all else fails. Another alternative is to reset the PRAM or the NVRAM by booting into open firmware ( command-option-o-f) and typing "reset-nvram (return) reset-all" without the "". This will also reset the PRAM.

[Patroclus]

don't want to sound stupid, but exactly what are kernel panics?

[ALG]

well looking at your output yours looks like it might be related to a usb driver/device.
A kernel panic usually is the result of two seperate applications trying to write to the same memory address. It also can be caused from 2 devices trying to use the same i/o address or irq.

[DPG4450Guy]

Definition:
1. What Unix does when a critical internal consistency checks fails in such a way that Unix cannot continue. The kernel attempts to print a short message on the console and write an image of memory into the swap area on disk. This can be analysed later using adb. The kernel will then either wait in a tight loop until the machine is rebooted or will initiate an automatic reboot.
Unix manual page: panic(8).
2. Action taken by software which discovers some fatal problem which prevents it from continuing to run.
Anyway, it is mostly represented by the "black curtain" falling down your machine's screen with a multi-language msg that you're screwed, and restart the computer. It is also seen as a never-ending spin of the beach ball . . .

[Patroclus]

ok thanks

[pkeene]

Thanks ALG. All quiet for the moment so maybe things have shaken themselves out.
Peter

[pkeene]

Thanks for the info Worf. No panics today so far, so I'm not going to touch anything for the moment. I was under the impression that fsck couldn't be done on a journaled drive. Anyway, If it happens again I'll give that and resetting PRAM a try.
Peter