November 2003 - Darwin-kernel

getsockname(.....) not working
by Matt Jaffa 21 Nov '03

21 Nov '03

Hi, I am developing a kernel Extension and right now I have the socket.h file included in my KEXT but it wont let me use getsockname function found in socket.h any reasons for this? It says it can't find the identifier. Why I need this is I overided the system table entry for the sendto system call, and in that call it gives you the integer socket identifier and I am trying to get who sent the sendto command and also what protocol and any ports it wants to use, etc. Any way of getting this info from overiding the sendto system table function? 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

Re: kernel related MPI problem?
by Justin Walker 21 Nov '03

21 Nov '03

On Friday, November 21, 2003, at 01:56 AM, Peter Klaver wrote: A collegue in my lab, knowledgeable in unix but not specifically in Darwin, wants to run MPI code on a dual G5, but there some problems when using the smp-compiled libraries (-comm=share). During testing the smp-MPICH software the following error occurs: P4xxxxx shmget error -1 As with most system calls, the return value '-1' is an indicator that something went wrong. It doesn't tell you what went wrong. For that information, refer to the global variable 'errno', which will (should) contain a non-zero value described in /usr/include/sys/errno.h. You can print a text string describing the error with "perror()" (see 'man perror' for details. Without that information, we'd just be shooting in the dark (or taking way too close a look at your source ;=}). Regards, 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 *--------------------------------------*-------------------------------* _______________________________________________ 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

kernel related MPI problem?
by Peter Klaver 21 Nov '03

21 Nov '03

Hi everyone, A collegue in my lab, knowledgeable in unix but not specifically in Darwin, wants to run MPI code on a dual G5, but there some problems when using the smp-compiled libraries (-comm=share). He suspects it's a kernel-related matter, see below. From his description, can anyone see if this is indeed the case? If anyone on this list happens to have smp-MPI running ok, then any other relevant info is of course also much welcomed. kind greets, Peter Shared memory problem. During testing the smp-MPICH software the following error occurs: P4xxxxx shmget error -1 This looks like a shmget error. Testing the operating system with: sysctl -a |grep sysv kern.sysv.shmmax: 4194304 kern.sysv.shmmin: 1 kern.sysv.shmmni: 32 kern.sysv.shmseg: 8 kern.sysv.shmall: 1024 Looks allright to me. I have tested the shmget call by the following C-program. #include <sys/types.h> #include <sys/ipc.h> #include <sys/shm.h> int main() { key_t key; int size; int shmflg; int shmid; size = 100; key = IPC_PRIVATE; shmflg = IPC_CREAT & IPC_EXCL & (7+7*8+7*64); shmflg = shmflg&IPC_CREAT; shmid=shmget(key, size, shmflg); printf("result %d/ \n",shmid); return 1; } It always returns -1. Runnig the C-program on other (non G5) platforms returns a normal shmid. my hardware platform is dual-G5 my OS version is 10.3 or 10.2.8, similar problems on both. -- +++++++++++++++++++++++++++++++++++++++ The requirements said: Windows 2000 or better. So I got a Macintosh. Peter Klaver private: Noordeinde 28 2311 CE Leiden 071-5137357 at work: Delft University of Technology e-mail T.P.C.Klaver(a)tnw.tudelft.nl room 222, 015-2789518 Rotterdamse Weg 137 2628 AL Delft the Netherlands _______________________________________________ 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: Accessing non-exported kernel symbols
by Vincent Lubet 21 Nov '03

21 Nov '03

On Nov 20, 2003, at 4:17 PM, Rob McKeever wrote: I'm wondering if there is any way to get access to non-exported kernel symbols from an IOKit kext. The question comes up after trying to find a way to access ubc_umount() for an experiment, but I'm sure there are other symbols I'll want to torture as time goes on. Before the obvious response comes back that if something isn't exported, I shouldn't be playing with it, I am aware of that there isn't any guarantee of compatibility between versions of non-exported symbols and that futzing around inside dark areas of the system could have dire consequences. This is primarily for my own use in exploring how the system works and what is possible. Another obvious response is to build your own version of xnu where you add the symbol to the export set. Vincent _______________________________________________ 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

Accessing non-exported kernel symbols
by Rob McKeever 21 Nov '03

21 Nov '03

I'm wondering if there is any way to get access to non-exported kernel symbols from an IOKit kext. The question comes up after trying to find a way to access ubc_umount() for an experiment, but I'm sure there are other symbols I'll want to torture as time goes on. Before the obvious response comes back that if something isn't exported, I shouldn't be playing with it, I am aware of that there isn't any guarantee of compatibility between versions of non-exported symbols and that futzing around inside dark areas of the system could have dire consequences. This is primarily for my own use in exploring how the system works and what is possible. Thanks in advance, -Rob -- /* drunk.. fix later */ _______________________________________________ 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: [Split off] Re: Kernel Extensions
by Matt Jaffa 19 Nov '03

19 Nov '03

Hi all, I included this in my kernel KEXT #include <sys/systm.h> and it is giving me 3 errors, because it systm.h included kernel/kernel.h and that seems to be buggy, whenever kernel.h hits a u_int any fixes? suggestions? 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

Re: Oddball VFS buffer question
by Paul Ripke 18 Nov '03

18 Nov '03

On Wednesday, Nov 19, 2003, at 06:49 Australia/Sydney, Rob McKeever wrote: Is there an easy, albeit unsupported way, to disable the VFS buffer cache for a given volume or mount point? I realize that any solution to this question is likely to not survive between kernel updates and could result in very bad things. Ideas? If support isn't there already (haven't looked at recent xnu), I would've thought it wouldn't be too hard to add, given support for F_NOCACHE in fcntl(2). I can see how something like the AIX 5.1 "mount -o dio ..." could be handy. We're using it for Oracle DBs, it makes quite a difference. With Xserve and Xserve RAID, this "feature" may be somewhat desirable. Cheers, -- Paul Ripke Unix/OpenVMS/TSM/DBA I love deadlines. I like the whooshing sound they make as they fly by. -- Douglas Adams _______________________________________________ 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

[Split off] Re: Kernel Extensions
by Peter Bierman 18 Nov '03

18 Nov '03

At 12:00 PM +0100 11/18/03, Stiphane Sudre wrote: On Monday, November 17, 2003, at 11:43 PM, Jim Magee wrote: And this brings up back to the suggestion that you use ipfw - because it was specifically designed to reflect this kind of traffic out to user-space. That is, if you are going to use anything on these machines at all (instead of just using a proxy server on your network as others have suggested). The argument that any admin user could change the firewall rules doesn't really hold water. They can remove your kext as well. This is why I'm still wondering why a kext needs to be root:wheel 644/755 and not just root:admin 644/755 when any admin user can be root:wheel if he wants and when he wants. I still don't understand this modification introduced in 10.2. Admins are allowed to become root as a matter of policy, not equivalency. This is the default policy provided by Apple, but end users and sysadmins are able to change this policy if they like. Most of the mechanisms that elevate admins to root go through the Security framework, which regulates this via the /etc/authorization file. The ones that don't are probably bugs. So to answer your question, the reason is to allow the distinction to be made, even if the default config doesn't make it. -pmb _______________________________________________ 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

Oddball VFS buffer question
by Rob McKeever 18 Nov '03

18 Nov '03

Is there an easy, albeit unsupported way, to disable the VFS buffer cache for a given volume or mount point? I realize that any solution to this question is likely to not survive between kernel updates and could result in very bad things. Ideas? -Rob -- "Quidquid latine dictum sit, altum viditur." _______________________________________________ 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: [Split off] Re: Kernel Extensions
by Stéphane Sudre 18 Nov '03

18 Nov '03

You're right but the problem was this feature was added for security reason because some kext were released with casual user's privileges. But sometimes ago in a not so far away galaxy, Installer.app was not able to install root:wheel items outside /System/Library/Extensions. So on one hand, there was a need for unnecessary too high security: (root:wheel vs root:admin) and on the other side, there was no obvious mean to respect it with the provided tools. On Tuesday, November 18, 2003, at 06:32 PM, Corey O'Connor wrote: Correct me if I'm wrong but... Even though any admin can be root:wheel if they want, they still have to su or sudo to do so. In my opinion this is a safety net against accidental kext modification. As with any mucking about in the kernel level, you needa be really sure what your doing. So I have no problems with more protections against modifying kext's. _______________________________________________ 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