Help with a kernel panic |
September, 23, 2002 8:49 PM |
corndog |
I believe I copied everything correctly. Here we go. BTW this is the error that I get when trying to install 10.2. I already have 10.1.5 running just fine. No debugger configured - dumping debug information MSR-00001030 Lastest stack backtrace for cpu 0: Backtrace: 0x00085936 0x00085F00 0x0008F648 0x00092688 Processing back via exception chain: Exception state (sv=oDCCEA00) PC=0x0e1ef1b8; MSR=0x00049030; DAR=0x000BB5000; DSISR=0x40000000; LR=0X0E1EF1B8; R1=0x076ABD20; XCP=0x00000008 (0 x 200 - Machine check) Backtrace: 0x0024D8A4 0x0E08CC00 0x002260B8 0x0003EE98 0x0003EDA0 Kernel loadable modules in backtrace (with dependencies): com.apple.iokit.IOAudioFamily (1.3.0b10)@0xe086000 Exception state (sv=0x0140EA00) PC=0x00000000;MSR=0x0000D030;DAR=0x00000000;DSIR=0x00000000;LR=0x00000000;R1=0x00000000;XCP=0x00000000(unknown) Kernel version: Darwin kernel version 6.0: Sat Jul 27 13:18:52 PDT 2002; Root:xnu/xnu-334.obj~1/RELEASE_PPC Panic: We are hanging here.... |
. |
RE: Help with a kernel panic |
September, 28, 2002 4:14 PM |
Tony.Scaminaci |
. |
I received the exact same error today when attempting to archive and install from 10.1.5 to 10.2. Turned out that the OS X installer corrupted the drive's directory sometime during CD #1. When the system rebooted to ask for CD #2, the kernel panicked with the debugger error. I force-rebooted into OS 9 and restarted the CD install process. Instead of rerunning the installer, I ran the 10.2 disk utility which found and fixed major directory damage. I also fixed permissions while I was at it. Finally, rebooted back into OS 9 and started up from the incomplete OS X drive. The system booted up correctly and then asked for the second CD as expected. Everything was fine from that point on. The 10.2.1 updater also caused minor directory damage during it's install, so I had to use OS X's DFA again to fix it and the permissions. Apple's installer really messes up the drives during install, and I took great care to run Techtool Pro on the 10.1.5 partition before running the Jaguar installer. |
. |
RE: Help with a kernel panic |
September, 28, 2002 10:06 AM |
corndog |
. |
I haven't tried installing in safe mode yet. I was going to try to install it on another internal drive and see if that would do it. Interestingly, booting back into 10.1.5 this morning, iTunes no longer works. Double clicking on a song get me no sound. The progress bar doesn't move, time counter doesn't more, nothing. It isn't actually locking up or freezing it just doesn't want to play any music. Updating to 3.0.1 didn't alleviate the problem. Unsure if this was caused by simply unplugging my speakers. I will try the install on another drive in safe mode this weekend and keep you posted. Thanks again. |
. |
RE: Help with a kernel panic |
September, 28, 2002 9:58 AM |
OSXGuru |
. |
One other thing you could try booting in "safe mode." To do this, you would hold down the "shift" key as XPostFacto restarts your computer. What this will do is prevent the audio-related kernel extensions from loading at all. You won't get any sound, but you also won't get a kernel panic. And I believe that enough of the kernel extensions should load to permit you to do the install. You can then upgrade to Mac OS X 10.2.1 and see if that helps. One possible difficulty is that I believe the ethernet kernel extensions may not load in a "Safe Boot", so you may not get any networking. Probably the best way to deal with this would be to install 10.2, and then move the IOAudioFamily.kext out of the way, and then reboot without using "Safe Mode". That might just do the trick. I suppose that one problem is that if you can't end up getting Mac OS X 10.2.1 to work to your satisfaction, it may be time-consuming to go back to 10.1.5. There would be a couple of other ways to deal with this, if the "safe boot" idea doesn't sound attractive. I could compile a custom version of XPostFacto that disables the audio (basically, another way of avoiding having the IOAudioFamily.kext load). Or there would be a way of disabling the audio by booting into the Open Firmware shell--basically, you would add a property named "AAPL,ignore" to the awacs node in the device tree. But the exact commands to do that are a little hard to remember--I would have to try it myself and write it down for you. |
. |
No Luck |
September, 28, 2002 12:09 AM |
corndog |
. |
Unplugging the speakers didn't help. I still get the kernel panic with the reference to IOAudioFamily. I can't complain about trying to put a V-12 into a Model T PM9500. Thanks for the input and keep up the good work. Have a good weekend. |
. |
RE: Help with a kernel panic |
September, 27, 2002 10:55 PM |
corndog |
. |
I do have external speakers hooked up to my 9500. I will disconnect and see where that takes me. I Am unsure how to take advantage of the new version of IOAudioFamily in 10.2.1 when I am dependent on the CD boot of 10.2. Any hoo, I will disconnect and see how it goes. Thanks for breaking down all of the Latin into something this mortal can understand. "Just unplug your speakers." Thanks for the help. I will keep you posted. |
. |
RE: Help with a kernel panic |
September, 27, 2002 4:25 PM |
OSXGuru |
. |
It looks as though this is a bug in the IOAudioDevice class (in the IOAudioFamily kernel extension). The interesting part of the backtrace looks like this: IOTimerEventSource::timeout(void*) + 120 IOAudioDevice::removeTimerEvent(OSObject*) + 212 OSCollectionIterator::getNextObject() + 36 So it looks as though a timeout is being generated, a timer event is being removed, and there is a bug in the process of removing it. It is possible that the code isn't expecting a timeout at this point--perhaps the timer has already been removed? There is some code in the removeTimerEvent method that cancels timeouts, so perhaps there is a race condition (if the timeout fires at exactly the wrong moment, perhaps that causes the problem). But this is fairly speculative--in any event, it appears to be a bug. One thing you could try is adjusting anything about your audio setup--if you have speakers plugged in, try removing them, if you don't, try plugging something in. It is possible that changing the hardware setup would avoid triggering the bug. Also, it is worth noting that there was a new version of IOAudioFamily for 10.2.1, so it is possible that this is a bug that got fixed (but I don't know that). The other possibility would be to compile a custom version of IOAudioFamily to try to narrow down exactly what the problem is. I could do that, in theory, assuming that playing with the audio hardware doesn't help. |