January 2003 - Darwin-kernel

Re: KEXT Debugging
by Andrew Gallatin 06 Jan '03

06 Jan '03

Justin C. Walker writes: > > FWIW, libkvm is present on Mac OS X/Darwin. Don't know if the Darwin > gdb is set up to use it though. Cool. I had no idea. Unfortunately, gdb doesn't appear to know about it though: % gdb GNU gdb 5.1-20020408 (Apple version gdb-228) (Sun Jul 14 10:07:24 GMT 2002) Copyright 2002 Free Software Foundation, Inc. <...> (gdb) help target <..> List of target subcommands: target async -- Use a remote computer via a serial line target child -- Unix child process (started by the "run" command) target cisco -- Use a remote machine via TCP target core-macho -- Use a core file as a target target exec -- Use an executable file as a target target extended-async -- Use a remote computer via a serial line target extended-remote -- Use a remote computer via a serial line target macos-child -- Mac OS X child process (started by the "run"command) target macos-exec -- Mac OS X executable target remote -- Use a remote computer via a serial line target remote-kdp -- Remotely debug a Mac OS X system using KDP <..> I'd expect a kcore target. Oh well.. 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: KEXT Debugging
by Alfonso Guerra 06 Jan '03

06 Jan '03

On Monday, January 6, 2003, at 04:08 PM, Andrew Gallatin wrote: However, you needn't debug on another mac. You can pull the gdb source from cvs and build a gdb which will work on another unix box. I debug on FreeBSD. I'd be happy to share the gdb binary... I've left it at http://people.freebsd.org/~gallatin/macgdb.gz Is it possible to debug an x86 from a mac? Alfonso _______________________________________________ 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: KEXT Debugging
by Eric Long 06 Jan '03

06 Jan '03

>> --Is that a dumb question? (-: > > Yup :-}. Hwh, heh. Well, I have quota to keep. (-; Thanks for the feedback. Eric _______________________________________________ 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: KEXT Debugging
by Justin C. Walker 06 Jan '03

06 Jan '03

On Monday, January 6, 2003, at 02:08 PM, Andrew Gallatin wrote: Eric Long writes: [snip] I haven't tried debugging in the kernel before. I found a document called "KernelProgramming.pdf" that provided a lot of instructions, which I've looked at, but it wasn't totally clear on these points. It only talks about remote debugging, but doesn't explicitly say that local debugging is out. If a breakpoint brings the whole kernel to a stop, then it makes sense that debugging would have to be done remotely only. Not entirely. If OSX had a libkvm and gdb support for it, you could at least examine variables locally while the machine is live. This can be very useful. As it is now, you must halt the machine and attach a remote debugger. FWIW, libkvm is present on Mac OS X/Darwin. Don't know if the Darwin gdb is set up to use it though. Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large * Institute for General Semantics | Men are from Earth. | Women are from Earth. | Deal with it. *--------------------------------------*-------------------------------* _______________________________________________ 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: KEXT Debugging
by Andrew Gallatin 06 Jan '03

06 Jan '03

Eric Long writes: > In debugging a network kernel extension, can I use gdb, or do I have to use > ddb? I'm manually loading and unloading the kext. It seems like I should > be able to use gdb. You can use gdb. I think using ddb requires a kernel rebuild. > Is it required that I debug remotely, or can I debug on the same machine the > extension is running on? --Is that a dumb question? (-: Yes, you must debug remotely. Local postmortem debugging is impossible, since OSX doesn't use a swap partition, and doesn't have the (highly useful) concept of crashdumps. However, you needn't debug on another mac. You can pull the gdb source from cvs and build a gdb which will work on another unix box. I debug on FreeBSD. I'd be happy to share the gdb binary... I've left it at http://people.freebsd.org/~gallatin/macgdb.gz Read http://www.opensource.apple.com/projects/documentation/howto/html/kext_tuto… to get started debugging. > I haven't tried debugging in the kernel before. I found a document called > "KernelProgramming.pdf" that provided a lot of instructions, which I've > looked at, but it wasn't totally clear on these points. It only talks about > remote debugging, but doesn't explicitly say that local debugging is out. > If a breakpoint brings the whole kernel to a stop, then it makes sense that > debugging would have to be done remotely only. Not entirely. If OSX had a libkvm and gdb support for it, you could at least examine variables locally while the machine is live. This can be very useful. As it is now, you must halt the machine and attach a remote debugger. 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: KEXT Debugging
by Justin C. Walker 06 Jan '03

06 Jan '03

On Monday, January 6, 2003, at 01:52 PM, Eric Long wrote: In debugging a network kernel extension, can I use gdb, or do I have to use ddb? I'm manually loading and unloading the kext. It seems like I should be able to use gdb. gdb Is it required that I debug remotely, or can I debug on the same machine the extension is running on? Yes. Think about hitting a breakpoint :-}. --Is that a dumb question? (-: Yup :-}. I haven't tried debugging in the kernel before. I found a document called "KernelProgramming.pdf" that provided a lot of instructions, which I've looked at, but it wasn't totally clear on these points. It only talks about remote debugging, but doesn't explicitly say that local debugging is out. If a breakpoint brings the whole kernel to a stop, then it makes sense that debugging would have to be done remotely only. It's out in general. I've used a debugger to peruse local system memory, although not on Mac OS X. Single-stepping and breakpoints in the kernel are hard to manage from the local system. I suppose one could think of ways to do it, but it might not be doing what you want (you'd have to, in some way, debug a virtual image of the kernel, not the kernel). All in all, you get fewer headaches by using remote debugging Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large * Institute for General Semantics | When LuteFisk is outlawed | Only outlaws will have | LuteFisk *--------------------------------------*-------------------------------* _______________________________________________ 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

KEXT Debugging
by Eric Long 06 Jan '03

06 Jan '03

In debugging a network kernel extension, can I use gdb, or do I have to use ddb? I'm manually loading and unloading the kext. It seems like I should be able to use gdb. Is it required that I debug remotely, or can I debug on the same machine the extension is running on? --Is that a dumb question? (-: I haven't tried debugging in the kernel before. I found a document called "KernelProgramming.pdf" that provided a lot of instructions, which I've looked at, but it wasn't totally clear on these points. It only talks about remote debugging, but doesn't explicitly say that local debugging is out. If a breakpoint brings the whole kernel to a stop, then it makes sense that debugging would have to be done remotely only. Thanks in advance for comments, Eric Long _______________________________________________ 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: Kext Info
by Dean Reece 06 Jan '03

06 Jan '03

Deepak, I think you want "systm.h" rather than "system.h". The odd spelling trips up many first-time kext developers, but it is historical so we haven't "corrected" it. Cheers, - Dean On Sunday, January 5, 2003, at 08:16 PM, Deepak Kumar Singh wrote: Hi , I tried to build Hello kernel extension "kext" on MAC OS 10.2 with July 2002 Project Builder.which is documented in this site http://developer.apple.com/techpubs/macosx/Darwin/howto/ kext_tutorials/hello _kext/hello_kext.html <http://developer.apple.com/techpubs/macosx/Darwin/howto/ kext_tutorials/hell o_kext/hello_kext.html> but it is giving error <sys/system.h> file not found ... though i included Kernel.framework and in header search path , i also mentioned the sys/system.h path.... but still i am getting the same error ..... Please help me out ... Many thanx in advance Regards Deepak _______________________________________________ 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

Kext Info
by Deepak Kumar Singh 06 Jan '03

06 Jan '03

Hi , I tried to build Hello kernel extension "kext" on MAC OS 10.2 with July 2002 Project Builder.which is documented in this site http://developer.apple.com/techpubs/macosx/Darwin/howto/kext_tutorials/hello _kext/hello_kext.html <http://developer.apple.com/techpubs/macosx/Darwin/howto/kext_tutorials/hell o_kext/hello_kext.html> but it is giving error <sys/system.h> file not found ... though i included Kernel.framework and in header search path , i also mentioned the sys/system.h path.... but still i am getting the same error ..... Please help me out ... Many thanx in advance Regards Deepak _______________________________________________ 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: socketvar.h
by Brian Bergstrand 03 Jan '03

03 Jan '03

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Friday, January 3, 2003, at 04:14 PM, Eric Long wrote: >> sockbuf is a kernel only structure, as should be evidenced by the >> KERNEL define wrapper. You can't expect to define KERNEL and have >> things work. Secondly, sockbuf is declared in the scope of the socket >> structure. >> >> IOW, if you aren't writing a kernel extension, don't even attempt to >> use the sockbuf (or socket) structure. > > I am writing a kernel extension. The #ifdef KERNEL thing is comes > from what > I saw in the source example I am drawing from. > Okay, sorry. The KERNEL define is unusual and it threw me off into thinking you were writing something in userland and trying to force a kernel structure to work. Usually, KERNEL is defined by PB or your Makefile. > Regarding scope... > > This is the part that is a bit confusing to me. I never define a > struct > within a struct like this header does. > > How should a forward declaration for this be declared? > > I am able to work with the declare sockbuf* paramaeters in C > implementation > files with no complications. I don't know what the compiler is > looking for > when I try to do the same in a C++ file. > Well, it probably has to do with C++ treating structs as classes (with implicit public access instead of implicit private access). So in C++, the sockbuf struct (class) is declared in the socket struct namespace and will only work from within that namespace. C has one namespace (well, really two) and so everything is global in C. This is not true in C++. Maybe something like "socket::sockbuf" would work? Don't know for sure. HTH. > > Eric Long > Brian Bergstrand <http://www.classicalguitar.net/brian/> PGP Key ID: 0xB6C7B6A2 My code has no bugs, only random new features. -----BEGIN PGP SIGNATURE----- Version: PGP 8.0 iQA/AwUBPhYBCXnR2Fu2x7aiEQILggCg3FLrDm6iCdH8LL3A1sr4Ep59WTQAoP9K 8L/WeOYrWEL9gwzKjOsIim4V =26b4 -----END PGP SIGNATURE----- _______________________________________________ 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