December 2003 - Darwin-kernel

char * causing kernel panics
by Matt Jaffa 16 Dec '03

16 Dec '03

Hi, Never mind where ever i declare a char * or caddr_t vairable it is causing a kernel Panic when i loaded a compiled version of my KEXT, despite the fact that there are other places in the code that have these variables declared. I took example project from nke and modified it. Now why would adding char * values cause it to panic? Thanks, Matt _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

char * varname causing kernel panics
by Matt Jaffa 16 Dec '03

16 Dec '03

Hi, I created a NKE using a dlil filter for filtering outgoing packet purposes. Within one of my functions, out_filtr(...,....,....,...,..,...) Function name or contents really don't matter. The problem I am having is that when ever i declare a char * varname; within this function and also as a global variable it causes the kernel panic. My intentions were not to have a global char * variable but that was for testing purposes. But I have declared a char * variable elsewhere in a different function of my code and it didn't cause a kernel panic. Any ideas? Matt _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: fsck vs. Journaled HFS+
by James Bucanek 15 Dec '03

15 Dec '03

>Date: Mon, 1 Dec 2003 18:25:58 +0000 >From: Quinn <eskimo1(a)apple.com> > >At 16:01 -0700 23/11/03, James Bucanek wrote: >>I'm trying to find some definitive word about using fsck on a >>Journaled HFS+ file system. I'm technical editor for a book on OS X >>and I'd like to get to the bottom of this. > >I forwarded your questions to the guy who implemented journaling in >HFS Plus and he just got back to me with a response... <clip> Thanks, Quinn, for taking the time to forward this to your contact on the HFS team, and for their comprehensive reply. Together with an excellent technical article in the Macintosh Developer's Journal, I think I've got all the information I was looking for. Except for one last nagging question. Looking at the boot logs, I'm pretty sure I know the answer but I'm still looking for a definitive statement. When you boot OS X in single user mode, the root filesystem is mounted read-only. When a journalled HFS volume is mounted read-only, is the journal still replayed? From my casual testing, it appears that it is -- despite the fact that this sort of goes against the concept of "read-only". If anyone has any information on the behavior of an HFS journalled filesystem when mounted read-only, I'm dying to find out. ______________________________________________________ James Bucanek <mailto:privatereply@gloaming.com> _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: 2 debugging questions
by Andrew Gallatin 15 Dec '03

15 Dec '03

Andrew Gallatin writes: > 1) How do I see the contents of the message buffer from a remote > gdb session?. The scenario is that the host has crashed, my > driver has spewed a few IOLogs, but I have no way to run > "dmesg" to see them. > > Is my only option to log in as ">console" and hook up a > monitor to the video card? To answer my own question, after following the twisty maze of _doprnt() callers, I finally figured out that its a more or less standard BSD msgbuf. This seems to work: (gdb) x/s msgbufp->msg_bufc 0x34c00c <temp_msgbuf+12>: "standard timeslicing quantum is 10000 us\nvm_page_bootstrap: 124567 free pages\nmig_table_max_displ = 68\nIOKit Component Version 7.0:\nWed Sep 24 15:43:30 PDT 2003; root(rcbuilder):RELEASE_PPC/iokit/RELE"... (gdb) 0x34c0d4 <temp_msgbuf+212>: "ASE\n90 prelinked modules\nCopyright (c) 1982, 1986, 1989, 1991, 1993\n\tThe Regents of the University of California. All rights reserved.\n\nusing 1310 buffer headers and 1310 cluster IO buffer headers\nDAR"... <...> Drew _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: 2 debugging questions
by Andrew Gallatin 15 Dec '03

15 Dec '03

Marek Kozubal writes: > > > 2) Can I reboot the (crashed) box via remote gdb? How? > > Walking down the hall to reboot the box is good excercise, but only > > up to a point ;) > > Go to lists.apple.com and look at the archives for this list and do a > search on "Rebooting the victim machine using gdb, the answer". Thanks, but it doesn't work for me: (gdb) l mx_reboot 539 #include <IOKit/IOPlatformExpert.h> 540 void mx_reboot(void); 541 542 void 543 mx_reboot() 544 { 545 (*PE_halt_restart)(kPERestartCPU); 546 } 547 548 (gdb) call mx_reboot() warning: Cannot insert breakpoint 0; disabling it. Error accessing memory address 0xcaf3578: Unknown error: -1 Maybe this is because I'm using a cross-compiled gdb on a different platform. Drew _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Repost: Patching sysent
by Quinn 15 Dec '03

15 Dec '03

Greetings I posted this a while back and got a lot fewer responses than I expected. I'm trying to find out why developers are patching the kernel system call dispatch table (sysent). If your product patches sysent, can you please drop me a line explaining what your product does and why it patches sysent. Apple is currently filling out the design for the Kernel Programming Interfaces (KPI) that will provide binary compatibility for BSD-based kernel extensions in a future version of Mac OS X. As part of this process, we need to understand why folks are doing bad things (such as patching sysent) so that we can provide KPIs to cover this functionality. If you're patching sysent, it's vital that you get in touch with me so that I can capture your product's requirements. If you responded to my previous email, there's no need to respond again. S+E -- Quinn "The Eskimo!" <http://www.apple.com/developer/> Apple Developer Technical Support * Networking, Communications, Hardware _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Dirty vnodes
by Sam Vaughan 14 Dec '03

14 Dec '03

I raised a Radar bug (number 3494492) a while ago to bring attention to the fact that it would be very useful if we had a way of querying the Mach layer to find out if a vnode has cached and/or dirty pages. The implementation for 'cached' should be fairly simple, but 'dirty' might be harder, as in the process of checking each page, new ones could become dirty. Our best approximation to 'dirty' at the moment is to look at the v_clen field of the vnode and return true if it's non-zero. This is less than ideal because it only means that the vnode was dirty at some point in the past - it might since have been cleaned. We're hitting a performance issue as a result of this. In Final Cut Pro for example, if you've rendered some video there will be vnodes to large files that were once dirty. Whenever a sync occurs, we decide that these vnodes are dirty and call ubc_pushdirty_range() on them. The box then grinds to a halt for a few seconds as the Mach code iterates all the pages looking for dirty ones to flush. As a result, video playback stops for a few seconds at regular intervals. Is it feasible for us to maintain our own dirty state for our vnodes? How do we know when a flush has finished writing out all the dirty pages? It seems that the transition from clean to dirty would be easy to handle, but the transition back from dirty to clean would be non-trivial. Sam _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: 2 debugging questions
by Marek Kozubal 12 Dec '03

12 Dec '03

On Fri, 12 Dec 2003, Andrew Gallatin wrote: > 1) How do I see the contents of the message buffer from a remote > gdb session?. The scenario is that the host has crashed, my > driver has spewed a few IOLogs, but I have no way to run > "dmesg" to see them. If you look at the FireWire SDK, there's something called FireLog which immediately puts the syslog() type messages out over the firewire bus which you can that receive on another computer. The instructions on how to do it are in the FireWire SDK package. > 2) Can I reboot the (crashed) box via remote gdb? How? > Walking down the hall to reboot the box is good excercise, but only > up to a point ;) Go to lists.apple.com and look at the archives for this list and do a search on "Rebooting the victim machine using gdb, the answer". --- Marek Kozubal marek(a)portents.com _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

2 debugging questions
by Andrew Gallatin 12 Dec '03

12 Dec '03

1) How do I see the contents of the message buffer from a remote gdb session?. The scenario is that the host has crashed, my driver has spewed a few IOLogs, but I have no way to run "dmesg" to see them. Is my only option to log in as ">console" and hook up a monitor to the video card? 2) Can I reboot the (crashed) box via remote gdb? How? Walking down the hall to reboot the box is good excercise, but only up to a point ;) In case it matters, my crashbox is a headless G5 running 10.3.1 Thanks, Drew _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: Kernel questions: Syscalls, ABIs, Signals, and Threads
by Andrew Begel 12 Dec '03

12 Dec '03

On Dec 12, 2003, at 12:59 PM, Jim Magee wrote: They sit on top of pthreads. So, an NSThread can call pthread functions without worry. Whew. That's good to know. PS: The ABI questions are best answered by others. However, there is Mach-O ABI documentation on the http://developer.apple.com website that you should definitely take a look at. Yes, I've read through the Mach-O Runtime Architecture manual, but this only gives a discussion of Mach-O PPC runtime semantics, not Mach-O i386 semantics. Thanks for the all the help, Andrew _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0