Version 2.2.5b4 |
December, 28, 2002 8:10 PM |
OSXGuru |
Version 2.2.5b4 is available (at this URL): http://eshop.macsales.com/OSXCenter/XPostFacto/Download/XPostFacto2.2.5b4.sit It does a better job of generating the Open Firmware name for devices--particularly, for the Radeon 7000. There is one remaining problem that can arise with UltraTek-based ATA cards (for a discussion, see the thread on version 2.2.5b3). I think I know how to fix it, and should have another version out soon. |
. |
RE: Version 2.2.5b4 |
January, 20, 2003 12:08 PM |
OSXGuru |
. |
Here's what the backtrace decodes to. It looks like the problem is occuring as the kernel is iterating over the IORegistry entries, looking for the "AppleKeyswitch" driver (not sure what that does). Since deleting the Extensions.mkext file seems to work for you, I wonder whether the Extensions.mkext file is not being created properly in your case for some reason. One workaround would be to write a little shell script that would delete that file each time you start up (after Mac OS X automatically creates it). Or you could try creating it manually, to see whether you get any interesting error messages or the like. -- IORegistryIterator::getNextObjectFlat() + 240 IORegistryIterator::getNextObjectFlat() + 228 IORegistryIterator::getNextObjectRecursive() + 40 IOService::getExistingServices(OSDictionary*, unsigned long, unsigned long) + 244 IOService::getMatchingServices(OSDictionary*) + 48 IOSCSIPrimaryCommandsDevice::start(IOService*) + 744 IOSCSIPeripheralDeviceType05::start(IOService*) + 36 IOService::startCandidate(IOService*) + 120 IOService::probeCandidates(OSOrderedSet*) + 2100 IOService::doServiceMatch(unsigned long) + 456 IOService::doServiceMatch(unsigned long) + 456 _IOConfigThread::main(_IOConfigThread*) + 284 ioThreadStart + 60 ioThreadStart |
. |
RE: Version 2.2.5b4 |
December, 30, 2002 4:46 PM |
gchron |
. |
Well here is the full data from the kernel panic. I hope that this helps No debugger configured - Dumping debug information. MSR=00001030 Latest Stack backtrace for cpu 0: Backtrace: 0x000856CC 0x00069C90 0x000287A8 0x0008F588 0x00092698 Proceeding back via exeption chain: Exception state (sv 0x0180E780) PC=0x0021857C; MSR=0x00009030; DAR=0x7C641B8C; DSISR= 0x40000000; LR=0x00218570; R1=0x0B743A70; XCP=0x0000000C (0x300 - Data access) Backtrace: 0x00218570 0x00218614 0x0021FAC4 0x0021FB90 0x13015890 0x134927BC 0x0021E8E8 0x0021E6E4 0x0021F034 0x0021F690 0x0020FB10 0x0020FAD4 Kernel loadableModules in backtrace (with dependencies) com.apple.IOSCSIMultimediaCommandsDevice (1.2.2) 0x13487000 com.apple.IOKit.IODVDStorageFamily (1.2) 0x133B5000 com.apple.IOKit.IOSCSIBlockCommandDevice (1.2.2) 0x132AA000 com.apple.IOKit.IOCDStorageFamily (1.2) 0x13356000 com.apple.IOKit.IOSCSIArchitectureModelFamily (1.2.2) 0x13012000 com.apple.IOKit.IOStorageFamily (1.2.3) 0x1321B000 com.apple.IOKit.IOSCSIArchitectutreModelFamily (1.2.2) 0x13012000 Exception State (Sv=0x12F10280) State (sv 0x12F10280) PC=0x00000000; MSR=0x0000D030; DAR=0x00000000; DSISR= 0x00000000; LR=0x0000D030; R1=0x0000D030; XCP=0x0000D030; (Unknown) |
. |
RE: Version 2.2.5b4 |
December, 29, 2002 5:16 PM |
OSXGuru |
. |
gchron: That is an interesting set of symptoms. Here are a few observations that I can make. The excerpt you posted from the panic text suggests that the panic does have something to do with the DVD driver. But we would need more of the text (pretty much all of it) in order to narrow things down. Unfortunately, there isn't any better way to do this at the moment than to just copy it down (or take a picture, if you have a digital camera). On newer machines, Apple has a scheme for storing panic text in NVRAM (to be retrieved at the next successful boot). Some rainy day I will try to implement this for older machines as well. It is interesting that deleting the Extensions.mkext file helps with the problem. This file is recreated at each startup if you delete it. I suppose that one inelegant workaround would be for you to write a little shell script to automatically delete the file, and set things up so that it runs each time after the file is created. The Extensions.mkext file is meant to be a cache of the kernel extensions that might load early in the boot process. In theory, the presence of the cache shouldn't change anything, but I guess it does in your case--it isn't obvious why. With respect to your comment about your ATTO card and ACARD card being on the same bus, I think that we are probably using the word "bus" in different ways. They may be on the same PCI bus, but they can't be on the same SCSI bus (since they each create their own SCSI bus). |
. |
RE: Version 2.2.5b4 |
December, 29, 2002 12:06 PM |
gchron |
. |
The as previus post at b3 I had seen the following text: The new feature in version 2.2.5 is that XPostFacto is now aware of the "device number" and "function number" for PCI devices. This helps with certain SCSI cards which differentiate between SCSI busses by using different function numbers. Well I have an ACARD IDE card and a ATTO apple SCSI card on my 9500. These two card are shown in the same bus and not at different one as you said above. I do not know if you meen that though. |
. |
RE: Version 2.2.5b4 |
December, 29, 2002 11:59 AM |
gchron |
. |
To continue the previous message I have here some of the dumping data I can see after the devfs /dev. If you need more data I can see the message again and feed you with data. Kernel loadableModules in backtrace (with dependencies) com.apple.IOSCSIMultimediaCommandsDevice (1.2.2) com.apple.IOKit.IODVDStorageFamily (1.2) com.apple.IOKit.IOSCSIBlockCommandDevice (1.2.2) com.apple.IOKit.IOCDStorageFamily (1.2) com.apple.IOKit.IOSCSIArchitectureModelFamily (1.2.2) com.apple.IOKit.IOStorageFamily (1.2.3) com.apple.IOKit.IOSCSIArchitectutreModelFamily (1.2.2) Exception State (Sv=0x12F11000) |
. |
RE: Version 2.2.5b4 |
December, 29, 2002 7:37 AM |
gchron |
. |
Well here are my findings. With the older version The problem with my DVD and acard ATA 133 card was resolved when I placed the DVD in the second IDE BUS. But when I installed the new XPF (I installed all the extensions, BOOTX etc) the system does not hang during the DEVFS /DEV but after that I displ;aies a kernel panic. The system cannot boot any more wioth the DVD connected. The only way is to remove the Extensions.mkext and Extensions.kextcache files. Then after the restart I see the DVD with no problem but only until the next restart. Then I have to trash the files again. I just reinstalled the 10.2.3 on a second small SCSI hard disk I have and the problem does not seem to happen. Is there are other file that I can trash in order to bring the system in healthy state?? Why the system worked fine before the update of XPF?? (I even tried to go to the pevius version and the problem cannot be fixed) |
. |
RE: Version 2.2.5b4 |
December, 29, 2002 2:03 AM |
willschou1 |
. |
RE: Version 2.2.5b4 December, 28, 2002 10:52 PM dreibel . Ryan - I discovered an interesting development tonight with the newest XPostFacto. Remember when I said Jaguar broke the external SCSI port on my J700, which prevented me from using my LaCie /Yamaha SCSI CD/RW burner? Sorry but IMHO it has to be something else that is keeping you from using your CDRW drive. I'm been using a J700 with OSX for over 2 years and there have been plenty of problems to fix over this time. SCSI internal and external and my Yamaha external 6416 CDRW have always worked.Wish I could tell you what the issue is best of luck. |
. |
RE: Version 2.2.5b4 |
December, 28, 2002 10:52 PM |
dreibel |
. |
Ryan - I discovered an interesting development tonight with the newest XPostFacto. Remember when I said Jaguar broke the external SCSI port on my J700, which prevented me from using my LaCie /Yamaha SCSI CD/RW burner? Well, though it doesn't show up in Apple System Profiler, when I launch iTunes it now recognizes a Yamaha CD burner! I STILL can't get it to burn , though, I get a message saying that the drive or the burning software is missing. Putting in a blank disc also doesn't call up Disc Burner on the desktop, and a regular CD-ROM still won't mount. However, this is encouraging - just keep doing what you're doing ! :) |
. |
RE: Version 2.2.5b4 |
December, 28, 2002 10:42 PM |
OSXGuru |
. |
OK, I see the problem now. Basically, the changes I made to fix the UltraTek problem had the side-effect of breaking the fix I had put in for the ATTO problem. I should be able to get them both working tomorrow. |
. |
RE: Version 2.2.5b4 |
December, 28, 2002 9:31 PM |
OSXGuru |
. |
egonzales21, could you e-mail to me your XPostFacto Log file (from your Preferences Folder, in the System Folder)? I did make a change which could affect the way that the ATTO card is named, and presumably that is the problem, but I can't quite see why. Your log file might help me figure it out. |
. |
RE: Version 2.2.5b4 |
December, 28, 2002 9:22 PM |
egonzales21 |
. |
Ryan thanks for all the work today. I have been working all day trying to get an ATTO UL2D, Tempo Trio, and a Radeon 7000 to coexist. I am sure it has to do with the names as well as the size of the NVRAM. Good new and Bad news. Have used 2.2.5b4. It indeed has fixed the Radeon 7000, but in the process it changed the name of the ATTO card. I was able to boot into open firmware and the boot-device read pci1/ATTO,ExpressPCIProUL2D@D/@0:6 in my case. By now everyone knows it needs to read pci1/ ATTO,ExpressPCIProUL2D@D,1/@0:6. In OF I changed the boot-device to read pci1/@D,1/@0:6 and the output device to pci1/@F/ATY,RV100ad_A and this was saved. I proceeded to boot in 10.2.3 where I am posting this now. Of course I took out the Tempo Trio again to be able to do this. Time to put it back in and see what XPF 2.2.5b5 has in store. Thank you once again. It seems once one thing gets fixed we give you another 4 or 5 to take care of. Sonnet owes you a lot of money (My opinion) |
. |
And version 2.2.5b5 |
December, 28, 2002 9:01 PM |
OSXGuru |
. |
And now, version 2.2.5b5. It should help with the UltraTek-based cards. I'm making the assumption that all the affected cards have Open Firmware names that begin with "UltraTek". I'm not absolutely sure this is true, but it does simplify things considerably :-) Here's the URL: http://eshop.macsales.com/OSXCenter/XPostFacto/Download/XPostFacto2.2.5b5.sit |