January 2005 - Darwin-kernel

Re: strange panic, debugging help wanted..
by Godfrey van der Linden 10 Jan '05

10 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hope this helps. Godfrey On Dec 22, , at 6:03, Andrew Gallatin wrote: The crash behaviour itself is very strange. The guy on site (machines are 3K miles from me) told me one crashed machine was scrolling in an infinate loop (he watched it for 10 minutes) printing register dumps. The latest panic.log I've gotten is very strange, with the stack trace on the panic'ed CPU always in what appears to be trap handling code. Here is the only panic log I've been able to get: Tue Dec 21 18:08:48 2004 Kernel version: Darwin Kernel Version 7.4.0: Wed May 12 16:58:24 PDT 2004; root:xnu/xnu-517.7.7.obj~7/RELEASE_PPC Kernel version: Darwin Kernel Version 7.4.0: Wed May 12 16:58:24 PDT 2004; root:xnu/xnu-517.7.7.obj~7/RELEASE_PPC I have the feeling that this is more of a "collateral damage" sort of panic, and the real problem has corrupted some state badly enough that any information regarding the real panic or exception is lost. Am I correct? Any suggestions on how to debug this? I'm setting up kernel dumps now, but I don't have much hope that I'll get one. Thanks, Drew _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/gvdl%40apple.com This email sent to gvdl(a)apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… I'd like to see the assembly that this kernel is running at 0x2CAE8. I'd be willing to bet an offset from a NULL pointer is being taken and that is why you are panicing. I think the 'r1' panic is a red herring, the first exception state is 'PC=0x0002CAE8; MSR=0x00001030; DAR=0x000000D4; DSISR=0x40000000; LR=0x0002 CAD8; R1=0x0CC33DB0; XCP=0x0000000C (0x300 - Data access)' Indicates that the r1 is valid at the time that the panic is taken. Do you have a symbolled kernel for the version that is taking the panic. If you can find out what routine was passed a NULL pointer you may have a suspect. Finally when you see this sort of crash, that is a crash in the kernel but none of your code is in the backtrace, then I'd suspect a teardown race of some sort. You or somebody else may be zero-ing a pointer early, or perhaps you are using a freed data structure. When this happens you often trash the data that is being resused by someother parts of the system. Our shop runs an automated "tinderbox" where a number of machines re-load our driver and run a series of QA tests each time a developer commits something to our source tree. In the last few days, some changes (probably changes in our driver) have caused our G5s to start crashing randomly. Since the crashes are random, they are hard to track down to a specific commit. Unresolved kernel trap(cpu 1): 0x300 - Data access DAR=0x00000000000000D4 PC=0x0 00000000002CAE8 Latest crash info for cpu 1: Exception state (sv=0x1DD39000) PC=0x0002CAE8; MSR=0x00001030; DAR=0x000000D4; DSISR=0x40000000; LR=0x0002 CAD8; R1=0x0CC33DB0; XCP=0x0000000C (0x300 - Data access) Backtrace: 0x0002CAD8 0x0002C8A8 0x0002C870 Proceeding back via exception chain: Exception state (sv=0x1DD39000) previously dumped as "Latest" state. skipping... Exception state (sv=0x009D8500) PC=0x00000000; MSR=0x0000D030; DAR=0x00000000; DSISR=0x00000000; LR=0x0000 0000; R1=0x00000000; XCP=0x00000000 (Unknown) panic(cpu 1): 0x300 - Data access Latest stack backtrace for cpu 1: Backtrace: 0x0008391C 0x00083E00 0x0001EDA4 0x00090E60 0x0009426C Proceeding back via exception chain: Exception state (sv=0x1DD39000) PC=0x0002CAE8; MSR=0x00001030; DAR=0x000000D4; DSISR=0x40000000; LR=0x0002 CAD8; R1=0x0CC33DB0; XCP=0x0000000C (0x300 - Data access) Backtrace: 0x0002CAD8 0x0002C8A8 0x0002C870 Exception state (sv=0x009D8500) PC=0x00000000; MSR=0x0000D030; DAR=0x00000000; DSISR=0x00000000; LR=0x0000 0000; R1=0x00000000; XCP=0x00000000 (Unknown) This email sent to site_archiver(a)lists.apple.com

1 0

Re: network bridging support?
by Joshua Graessley 10 Jan '05

10 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Thanks, -josh On Jan 10, 2005, at 11:22 AM, Jonathan Buhacoff wrote: This email sent to jgraessley(a)apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com The lower layers of the kernel, where interfaces tie in to the stack on Darwin is very different from FreeBSD. Darwin does contain some of the FreeBSD bridging code in #ifdefs throughout the stack, but the code won't work. With enough effort it could probably be made to work. Feel free to file a bug. I'm not sure where this will end up on the priority list, but it won't even make the list if we don't have a bug to track it ;) Linux and FreeBSD have bridging support built into the kernels. I have Darwin 7.7 but I can't find any hints online on how to enable bridging with it. Can anyone point me in the right direction? Thanks, Jonathan, from Napa, California _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/jgraessley% 40apple.com smime.p7s

1 0

network bridging support?
by Jonathan Buhacoff 10 Jan '05

10 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… Linux and FreeBSD have bridging support built into the kernels. I have Darwin 7.7 but I can't find any hints online on how to enable bridging with it. Can anyone point me in the right direction? Thanks, Jonathan, from Napa, California This email sent to site_archiver(a)lists.apple.com

1 0

Re: Question on VFS-Lock
by Shawn Erickson 10 Jan '05

10 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Jan 9, 2005, at 10:06 PM, Yogesh P wrote: Hi Folks, I am working on the VFS-Lock patch available for linux which is use to get the functionality of locking and unlocking the filesystem. Is anyone is working on to get the same functionality on MAC OS X which makes the filesystem lock unlock???? -Shawn _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… Is darwin kernel has some kind of mechanism for locking or unlocking the filesystem??? Any pointers,reply certainly helps. Can you better online what you mean by lock/unlock. Locks out IO, Locks out fs changes, etc. What is the main goal for this product, what type of feature are you trying to allow? Note that on Mac OS X the swap space is made of files in the boot file system (by default) and it is dynamic, files can come and go as swap needs change. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Memory Regions in Mach 3.0
by Justin Walker 10 Jan '05

10 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Jan 10, 2005, at 5:28, Mr. K. D. Ritson wrote: I'm new here so please be gentle and I hope I'm in the right place. Cheers, Justin -- Justin C. Walker, Curmudgeon-At-Large * Institute for General Semantics | Some people have a mental | horizon of radius zero, and | call it their point of view. | -- David Hilbert *--------------------------------------*-------------------------------* _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… I'm writing a thesis on virtual memory memory management in the XNU/Mach3/Darwin kernel. Mach 3 supports sparse memory allocation by using "Memory Regions" and I'm looking for some information on the structure of such memory regions. Can anyone point me in the direction of some detailed online info in this direction? There may be some info on the Mach stuff on-line at CMU. For Darwin ('xnu' is the project name for the darwin kernel), there is the best possible documentation: the source :-}. See also the Darwin site on the Apple Developer website. Since you are writing a thesis, you might want to get familiar with 'google'. This email sent to site_archiver(a)lists.apple.com

1 0

Memory Regions in Mach 3.0
by Mr. K. D. Ritson 10 Jan '05

10 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Importance: Normal I'm new here so please be gentle and I hope I'm in the right place. I'm writing a thesis on virtual memory memory management in the XNU/Mach3/Darwin kernel. Mach 3 supports sparse memory allocation by using "Memory Regions" and I'm looking for some information on the structure of such memory regions. Can anyone point me in the direction of some detailed online info in this direction? I'm grateful for any help. Cheers. -k _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

Question on VFS-Lock
by Yogesh P 10 Jan '05

10 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hi Folks, I am working on the VFS-Lock patch available for linux which is use to get the functionality of locking and unlocking the filesystem. Is anyone is working on to get the same functionality on MAC OS X which makes the filesystem lock unlock???? Is darwin kernel has some kind of mechanism for locking or unlocking the filesystem??? Any pointers,reply certainly helps. Thanks, YOGESH PAHILWAN (SOFTWARE ENGINNER) S P Software Technologies (I) Pvt. Ltd. (An ISO 9001:2000 Company) Pune, Maharashtra, India - 411 004. Tel : 91-20-4006154, 4006983, 4007503. Fax : 91-20-4006153. Email : pahilwan.yogesh(a)spsoftindia.com Visit us at : www.spsoftindia.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Communicating from kernel extension to application?
by Brian Kendall 10 Jan '05

10 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Organization: Qrrbrbirlbel User-agent: Opera M2/7.54 (MacPPC, build 1840) TCPLogger is DTS sample code. If you search the developer support web page, it should turn up. You'll want to use a kernel control socket. TCPLogger sample should show you how. There is some documentation in the header file at /usr/include/sys/kern_control.h I still can't find TCPLogger. Searching the developer site only yields two results to documents mentioning the existence of TCPLogger, but nothing of TCPLogger itself. There must be something I'm missing here in terms of finding it. Where exactly is it? - Brian _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Communicating from kernel extension to application?
by Dean Reece 09 Jan '05

09 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Brian, Cheers, - Dean _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… On Jan 8, 2005, at 11:48 PM, Brian Kendall wrote: I'm wondering what methods exist for having a kernel extension send a message of some kind that can be received by an application in user space. I searched the archives, and apparently circa a couple of years ago someone asked a similar question and was directed to the source of an example kext called TCPLogger in the Darwin source, but it doesn't look like it exists anymore. Is there anywhere I can find that source, or forgetting that, how could one send a message from a kext to an application? What type of kext are you developing, or more specifically, what API sets are you already using? There are different approaches for IOKit kexts and BSD kexts, and variants within each of these depending on the details. Life will probably be easier if you stick to interfaces closer to what you are already using. Also, pretty much all kernel->user notification channels require the user-land piece to make the first move; signing up for a notification, getting a mach port and waiting for a message, select, etc...). This is done to avoid having kernel threads transition into a user-level task. Instead, the user-level task supplies a thread that is blocked in the kernel waiting for something to happen (synchronous), or the app establishes an event thread and signs it up for notifications (async). Examples include read()/write(), select, mach ports (which underly IOKit messaging), and even signals – all of which execute in a user-supplied context. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Communicating from kernel extension to application?
by Joshua Graessley 09 Jan '05

09 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com -josh On Jan 8, 2005, at 11:48 PM, Brian Kendall wrote: This email sent to jgraessley(a)apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com TCPLogger is DTS sample code. If you search the developer support web page, it should turn up. You'll want to use a kernel control socket. TCPLogger sample should show you how. There is some documentation in the header file at /usr/include/sys/kern_control.h I'm wondering what methods exist for having a kernel extension send a message of some kind that can be received by an application in user space. I searched the archives, and apparently circa a couple of years ago someone asked a similar question and was directed to the source of an example kext called TCPLogger in the Darwin source, but it doesn't look like it exists anymore. Is there anywhere I can find that source, or forgetting that, how could one send a message from a kext to an application? - Brian _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/jgraessley% 40apple.com smime.p7s

1 0