[chasg]

Hi All,

Well, my dual 2.0GHz (2nd gen) G5 is having some serious problems. After some annoying symptoms (lots of problems with Toast and failed burns, plus apps more and more frequently crashing), I now get a kernel panic within a minute or so of starting up my machine (even if I do nothing and wait for it to die).

I used to troubleshoot professionally back in the OS9 days, and some of those skills have translated on to OSX, but my machines have been running so flawlessly for so long, my skills probably need some assistance.

So, before I begin a troubleshooting blitz, I'd like to see if anyone can perhaps direct me to an area to address first.

The machine is running OSX 10.4.2, and has been since it came out (without the kernel panic problem). The RAM is what I installed when I got it (2.5 Gb, if I remember correctly), I have two internal drives (again, installed when I got it, and no indication of problems all this time) plus an external firewire800 drive (also a long time resident).

The only new software I've used recently is Aperture checker (to see if my machine will run Aperture, which I'm not going to get until version 2.5 at least!), everything else has been unchanged for ages.

So, I'm not sure what a kernel panic is a symptom of. Am I looking at failure of RAM, HD, the OS? I haven't even run Disk First Aid yet (though I did zap the PRAM out of reflex :-), not to mention DiskWarrior. I just want to do this fix in the most efficient way possible.

Anyone have any suggestions before I pull out the tools and begin taking this sucker apart? Thanks in advance for any help,

Chas

[seanc]

Post your panic log here and we'll see if we can identify the problem from it.

[tooki]

I'd certainly run DiskWarrior if you have it. It's not like it takes that long to do!

tooki

[Watson]

Try using the Apple Hardware Test, which should be on the Software Restore and Install DVD. Hopefully you will not have to grab a Torx 8.

Froom Shanka

[chasg]

Thanks for all the responses.

I'm presently running the extended Apple Hardware test (good suggestion Watson, I haven't used this util for ages and sortof forgot about it). It says a test will take 15 to 30 minutes, depending on the amount of memory. Heck, I'm at 40mins and the bar is only 2/3 of the way through (2.5Gb isn't all that much memory, is it?). At least I've got time to reply to this thread :-)

seanc: that's one of the things that's been worrying me. The useful "report to Apple" dialog comes up after I restart, if I hit "report" (I think that's what the button says), a report doesn't get generated. Where can I find the log file so I can post it here? (if necessary, I'm going to run Disk First Aid and Disk Warrior first).

tooki: can do, will do, thanks.

Watson: the torx 8 is anxious to come out of the toolbox, here's hoping it'll remain frustrated!

Chas

[SMacTech]

Take out some of your RAM and test for KPs. I have had bad RAM that never showed up in any Hardware test. I replaced it, and all my problems went away.

[seanc]

The log we're looking for in this case is panic.log

It's in:
[Your Harddrive]/Library/Logs/

[chasg]

Just an update.

I ran Disk First Aid from my original install discs (OSX 10.3), and it found some errors it couldn't fix (in the catalog, if I remember correctly). I dug out my Tiger install disc and the version of DFA managed to fix those probs. Didn't fix the KPs though, once I'd restarted.

I then used my Tiger disc to install a new version of Tiger, and archived the previous System. That didn't fix the problem (very strange, what could have been brought over to the new system that would cause the kernel panics to continue? (I don't think it's hardware that is the cause, see below).

I then installed Tiger on my external firewire disk (LaCie 500Mb), and took the risk of copying over everything during the install process ('cause I'm prob going to wipe the first disc, I don't want to run my machine from an external drive). No KPs so far, so I really don't think it's hardware related.

Given the evidence (plus log below), any chance that I can regain my internal disk by replacing some errant file in the afflicted System?

<grr> two days wasted! (still, much better than the OS9 days, 2 days wasted vs. months and months of uptime ain't bad!)

Chas

The panic log (only two entries):

Thu Dec 15 13:47:00 2005
panic(cpu 0 caller 0x46E5DAEC): DART entry exception: HyperTransport write logical page 0x05440

Latest stack backtrace for cpu 0:
Backtrace:
0x00095564 0x00095A7C 0x00026838 0x46E5DAEC 0x4F54AD70 0x4F54ADA8 0x4F50F048 0x002CDC8C
0x002CCB54 0x000A9654
Kernel loadable modules in backtrace (with dependencies):
com.apple.driver.AppleGPIO(1.1.8)@0x4f548000
dependency: com.apple.driver.IOPlatformFunction(1.7.0b1)@0x46e 47000
com.apple.driver.MacIOGPIO(1.1.8)@0x4f50d000
com.apple.driver.AppleMacRISC4PE(1.6.0d9)@0x46e570 00
dependency: com.apple.iokit.IOPCIFamily(1.6)@0x46e39000
dependency: com.apple.driver.IOPlatformFunction(1.7.0b1)@0x46e 47000
Proceeding back via exception chain:
Exception state (sv=0x00A1B280)
PC=0x00000000; MSR=0x0000D030; DAR=0x00000000; DSISR=0x00000000; LR=0x00000000; R1=0x00000000; XCP=0x00000000 (Unknown)

Kernel version:
Darwin Kernel Version 8.0.0: Sat Mar 26 14:15:22 PST 2005; root:xnu-792.obj~1/RELEASE_PPC
*********

Thu Dec 15 13:50:25 2005
panic(cpu 0 caller 0x00889AEC): DART entry exception: HyperTransport write logical page 0x08E40

Latest stack backtrace for cpu 0:
Backtrace:
0x00095564 0x00095A7C 0x00026838 0x00889AEC 0x007ECD70 0x007ECDA8 0x00655048 0x002CDC8C
0x002CCB54 0x000A9654
Kernel loadable modules in backtrace (with dependencies):
com.apple.driver.AppleGPIO(1.1.8)@0x7ea000
dependency: com.apple.driver.IOPlatformFunction(1.7.0b1)@0x4d0 000
com.apple.driver.MacIOGPIO(1.1.8)@0x653000
com.apple.driver.AppleMacRISC4PE(1.6.0d9)@0x883000
dependency: com.apple.iokit.IOPCIFamily(1.6)@0x452000
dependency: com.apple.driver.IOPlatformFunction(1.7.0b1)@0x4d0 000
Proceeding back via exception chain:
Exception state (sv=0x00DAF280)
PC=0x00000000; MSR=0x0000D030; DAR=0x00000000; DSISR=0x00000000; LR=0x00000000; R1=0x00000000; XCP=0x00000000 (Unknown)

Kernel version:
Darwin Kernel Version 8.0.0: Sat Mar 26 14:15:22 PST 2005; root:xnu-792.obj~1/RELEASE_PPC

*********

[astepanuks]

Hi,
Looks like one of the items in the backtrace is misbehaving on my Mac too.
Does anyone know what is com.apple.iokit.IOATAFamily.

Cheers
Andy


panic(cpu 0 caller 0x000A8D00): Uncorrectable machine check: pc = 000000000053DD0C, msr = 0000000000149030, dsisr = 42000000, dar = 00000000E0787000
AsyncSrc = 0000000000000000, CoreFIR = 0000000000000000
L2FIR = 0000000000000000, BusFir = 0000000000000000

Latest stack backtrace for cpu 0:
Backtrace:
0x00095698 0x00095BB0 0x0002683C 0x000A8D00 0x000A7F90 0x000ABC80
Proceeding back via exception chain:
Exception state (sv=0x34745000)
PC=0x0053DD0C; MSR=0x00149030; DAR=0xE0787000; DSISR=0x42000000; LR=0x00543308; R1=0x1F44BCD0; XCP=0x00000008 (0x200 - Machine check)
Backtrace:
0x002CF81C 0x00543308 0x0053D3B8 0x00543D24 0x0003C744 0x000A9814
Kernel loadable modules in backtrace (with dependencies):
com.apple.iokit.IOATAFamily(1.6.0f2)@0x537000
Exception state (sv=0x4A52D280)
PC=0x00000000; MSR=0x0000D030; DAR=0x00000000; DSISR=0x00000000; LR=0x00000000; R1=0x00000000; XCP=0x00000000 (Unknown)

[jamil5454]

My guess is the DVD drive or the internal hard drive. What happens when you Verify your disk?

[astepanuks]

Originally Posted by jamil5454

My guess is the DVD drive or the internal hard drive. What happens when you Verify your disk?

I am getting mixed messages here.
If I launch Disk Utility, only boot disk is SMART verfied and 2 other hard drives are reported that SMART is not supported.
However if I launch System profiler, it reports that all drives are SMART verified.

This panic started to happen only recently when I tweeked disk sleep in Energy panel. (I de selected sleep option whenever possible as it took forever for disks to spin up and was annoing).

AFter the tweek, Mac would crash with kernel panic every time after 3 or 4 minutes after it is woken from deep sleep.

Another recent development is that CDs or DVDs would not eject in first attempt. If I would eject the disk, it would dissaper from Finder and the would be remounted. Only then I can eject the disk. Weird. PS, it's a Pioneer 107 drive.

[Detrius]

Originally Posted by astepanuks

Hi,
Looks like one of the items in the backtrace is misbehaving on my Mac too.
Does anyone know what is com.apple.iokit.IOATAFamily.

Cheers
Andy
...

com.apple.iokit.IOATAFamily would be IO ATA Family in Apple's IOKit drivers. ATA is also called IDE, which is what all of your drives are running on.

I had a much longer response written up, but then I realized I was reading two different issues... You probably have a hardware problem, and it's likely the hard drive itself, but without more information, it's hard to narrow down what's going on.


As for chasg, I'm leaning towards logic board, though that could go many different ways as well. Don't rely on Apple Hardware Test. It tends to pretty easily pass faulty hardware. There is a program called Rember that can help you out with checking your RAM, but that's about it. TechTool Pro just blows as far as hardware tests are concerned... except hard drive hardware tests. It's good at that.

[astepanuks]

Thanks Detrius,
I will try to clone my boot drive to another one and try to replicate the problem.
Funny thing is that I am using SIIG SATA PCI card with 2 Maxtor SATA drive off it. One of these drives is my boot drive. Apple profiler reports this card as ATA interface (?) and both SATA drives as PATA.
I e-mailed SIIG with this reporting glitch and they are looking into it. However, Mike from xlr8yourmac recons that it is normal for Tiger no to report accurately on PCI SATA cards. Go figure
Cheers
Andy

[chasg]

Hi Detrius, thanks for the info, though a logic board fault is more than alarming. What makes you think the fault lies there? I'm going to have a pretty hard time convincing Apple to replace my logic board if I can't pinpoint the problem :-/

Andy, despite the superficial similiarities of our respective problems, it might have been a good idea to start your own thread so that we could both could retain focus from our contributors.

Chas

[Dr. Smoke]

Hi, Chas.

This type of panic generally indicates a that a device has attempted to perform a Direct Memory Access (DMA) read or write to an unprepared page. DMA involves accessing memory while bypassing the CPU, a function of all modern computer architectures. Hypertransport is a high-speed bus architecture between the computer's memory controller and its device I/O.

Specifically, a device has tried to read or write memory via DMA that has not been prepared by an IOMemoryDescriptor.

Read-related panics reveal themselves by beginning in the panic log:

Code:
panic(cpu 1 caller hex_address_1): DART entry exception: HyperTransport read logical page hex_address_2

Write-related panic begins

Code:
panic(cpu 1 caller  hex_address_1): DART entry exception: HyperTransport write logical page hex_address_2

where hex_address_n is a hexadecimal addresses, such as 0x00D4FAEC or 0x00C40.

The first Power Mac G5 desktops did not have an error facility for the DMA Address Relocation Table (DART). All later Power Mac G5 computers have such a facility. A kernel panic results when a PCI bus master device accesses memory at an address that is not mapped by the DART. Examples of PCI bus master devices include disk controllers, RAID controllers, and video cards employing DMA.

Generally this indicates an incompatible peripheral, often a PCI card, but incompatible AGP cards can also cause problems.

For additional technical details, see the following documentation corresponding to the G5-equipped Mac you are using:

- The "Bridge and Memory Controller" section of the "Architecture" chapter of the Apple Developer Connection document Power Mac G5 Developer Note.

For additional troubleshooting suggestions, see my Resolving Kernel Panics FAQ. This FAQ includes step-by-step instructions for identifying and resolving some of the most common causes of kernel panics