November 2001 - Darwin-kernel

Re: How to determime process name from pid?
by Chaskiel M Grundman 19 Nov '01

19 Nov '01

Based on reading ps sources: #include <sys/sysctl.h> int mib[4]; struct kinfo_proc p; int bufSize; mib[0]=CTL_KERN; mib[1]=KERN_PROC; mib[2]=KERN_PROC_PID; mib[3]=the_pid; bufSize=sizeof(p); error=sysctl(mib,4, &p, &bufSize, NULL, 0); if error is 0, the data you want is in p->ki_p->p_comm This is a non-privileged operation.

1 0

Re: Core99/Powermac3,1 support missing in xnu since Jan 2001?
by Chris Bednar 19 Nov '01

19 Nov '01

> Ah. An mkext is a multi-kext. You can actually build one using the > mkextcache tool. It contains all the kexts you need to boot stuffed into Well, yes, you can build one, and as you hint, /etc/rc does so at each bootup, which I learned to my chagrin a little before I got your note. It doesn't appear that mkextcache can produce a working Extensions.mkext for my system. It boots fine if I delete that file before I start up, so the extensions present are apparently OK (well, the ones I need, anyway). I do get one ``Error: Can't get bundle identifier.'' message when I run it; maybe that means something... It does report Compressing /System/Library/Extensions/AppleCore99PE.kext/Contents/Info.plist 2220 => 811 Compressing /System/Library/Extensions/AppleCore99PE.kext/Contents/MacOS/AppleCore99PE 89968 => 30842 which looks promising, but doesn't seem to work. I did move all extensions which appeared not to have been installed with 10.1, but that still didn't help. For me, the workaround is just to delete Extensions.mkext, and to comment out the bit in /etc/rc that generates it. The difference in boot time is fairly miniscule, and I don't reboot stable operating systems terribly often ;) Again, I think a lot of people are having this problem (check the message boards), so it might be a good idea to get something about this into the troubleshooting section online. If I find out which extension is causing trouble, I'll let everybody know... Thanks ---- Chris J. Bednar <http://optics.tamu.edu/~bednar/> Director, Distributed Computing Product Group http://AdvancedDataSolutions.com/

1 0

How to determime process name from pid?
by Kevin Bohan 19 Nov '01

19 Nov '01

If I have a PID is there a way (not is a shell script, but in C code) to determine what the name of the process whose PID this is, if there is a process with this PID. Kirby _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp

1 0

What is the preferred visual notification method from a daemon?
by Kevin Bohan 19 Nov '01

19 Nov '01

Hi, I am writing daemon that shuts down the system in response to an external event. The shutdown is performed by a daemon process. I don't really have a UI for this app, except for a pref pane to configure settings. I would like the user to be "notified" prior to shutdown. By notify, I mean a visual notification message/dialog pop-up, whatever. As I daemon I know I cannot directly show any UI elements. What is the preferred method of displaying a user-notification in this case. Kirby _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp

1 0

Re: Core99/Powermac3,1 support missing in xnu since Jan 2001?
by Louis Gerbarg 17 Nov '01

17 Nov '01

Ah. An mkext is a multi-kext. You can actually build one using the mkextcache tool. It contains all the kexts you need to boot stuffed into one file. The booter checks the dates of the mkext and the Extensions folder, and if the uses whichever is newer. Booting using an mkext is way faster than booting without one. The system startup scripts will build a new mkext if the current one is out of date. What that means is if you put a new kernel extension in /System/Library/Extensions the first time you reboot you will take a lot longer, and will rebuild a new mkext cache using the new extensions. Copying the mkext from the CD is not a garanteed solution. If it turns out something in your extensions folder is corrupt then when you add any new extensions that cause the mkext to be rebuilt you will build it with the corrupt extension. My advice to you to delete the mkext entirely and reboot. If you boot fine you will have an mkext that matches your system generated. If not you know you have to replace your Extensions folder. Louis On Friday, November 16, 2001, at 06:07 PM, Chris Bednar wrote: Yes, AppleCore99PE and AppleMacRISC2 support have been removed from the kernel proper. I think the cube should work with either. What version of Darwin are you trying to run. Those drivers should have been installed with Darwin 1.3.1 or 1.4.1. Are you trying to update an old 1.2.1 based system? It was a straight 10.0 setup, I think. Other xnu updates I had tried did the same thing to me as the 10.1 update did. Actually, the drivers were all installed. The file that was foobar was /System/Libraries/Externsions.mkext, which I guess is just a list of available kernel modules. The installer creates one that's about half the size of the one on the CD. After hunting around with ResEdit and Sherlock (WHY do you people think I need to be protected from myself with hidden files?) I copied that file from the CD to my disk, and it boots. I did try another time with the installer before I did all of this, just to be sure. Now, I just have to figure out how to get my user accounts straightened out again (I'm getting them from an NIS server) and should be good to go. My NFS mounts stayed as they were, and the NetInfo YP stuff is still there; if I remember right, I just need to edit a file somewhere to turn that back on. Anyway, the installer definitely has a bug for platforms that need kernel extensions to run, and I know a lot of people have had this problem, so I hope someone will pick up on this and fix it. By the way, now that I have the XFree86 crutch available, I think I'm just about ready to stop booting linuxppc on this thing... maybe our next big cluster will be Apple after all! ----- Chris J. Bednar <http://optics.tamu.edu/~bednar/> Director, Distributed Computing Product Group http://AdvancedDataSolutions.com/

1 0

Re: Core99/Powermac3,1 support missing in xnu since Jan 2001?
by Chris Bednar 17 Nov '01

17 Nov '01

> Yes, AppleCore99PE and AppleMacRISC2 support have been removed from the > kernel proper. I think the cube should work with either. > > What version of Darwin are you trying to run. Those drivers should have > been installed with Darwin 1.3.1 or 1.4.1. Are you trying to update an > old 1.2.1 based system? It was a straight 10.0 setup, I think. Other xnu updates I had tried did the same thing to me as the 10.1 update did. Actually, the drivers were all installed. The file that was foobar was /System/Libraries/Externsions.mkext, which I guess is just a list of available kernel modules. The installer creates one that's about half the size of the one on the CD. After hunting around with ResEdit and Sherlock (WHY do you people think I need to be protected from myself with hidden files?) I copied that file from the CD to my disk, and it boots. I did try another time with the installer before I did all of this, just to be sure. Now, I just have to figure out how to get my user accounts straightened out again (I'm getting them from an NIS server) and should be good to go. My NFS mounts stayed as they were, and the NetInfo YP stuff is still there; if I remember right, I just need to edit a file somewhere to turn that back on. Anyway, the installer definitely has a bug for platforms that need kernel extensions to run, and I know a lot of people have had this problem, so I hope someone will pick up on this and fix it. By the way, now that I have the XFree86 crutch available, I think I'm just about ready to stop booting linuxppc on this thing... maybe our next big cluster will be Apple after all! ----- Chris J. Bednar <http://optics.tamu.edu/~bednar/> Director, Distributed Computing Product Group http://AdvancedDataSolutions.com/

1 0

Re: Core99/Powermac3,1 support missing in xnu since Jan 2001?
by Louis Gerbarg 17 Nov '01

17 Nov '01

Yes, AppleCore99PE and AppleMacRISC2 support have been removed from the kernel proper. I think the cube should work with either. What version of Darwin are you trying to run. Those drivers should have been installed with Darwin 1.3.1 or 1.4.1. Are you trying to update an old 1.2.1 based system? Louis On Friday, November 16, 2001, at 12:24 PM, Chris Bednar wrote: Hi I haven't been able to boot any Darwin kernel newer than Jan 2001 on my AGP G4/400; it fails with the following panic: Unable to find driver for this platform: PowerMac3,1 Core99 support (PowerMac3,1, for example) seems to have dropped out of xnu on 4 Jan 2001 (cvs revision 1.1.1.21 of xnu/iokit/KernelConfigTables.cpp). Since then, I get a kernel panic when I try to boot. This happened first when I let OSX auto-update the first time I installed the official release. It's still happening with the OSX 10.1 upgrade. Do I need to do something special? Should it be finding a module somewhere? I see AppleCore99PE seems to have broken off into a separate project... Thanks ---- Chris J. Bednar <http://optics.tamu.edu/~bednar/> Director, Distributed Computing Product Group http://AdvancedDataSolutions.com/ _______________________________________________ darwinos-users mailing list darwinos-users(a)lists.apple.com http://www.lists.apple.com/mailman/listinfo/darwinos-users

1 0

Core99/Powermac3,1 support missing in xnu since Jan 2001?
by Chris Bednar 16 Nov '01

16 Nov '01

Hi I haven't been able to boot any Darwin kernel newer than Jan 2001 on my AGP G4/400; it fails with the following panic: Unable to find driver for this platform: PowerMac3,1 Core99 support (PowerMac3,1, for example) seems to have dropped out of xnu on 4 Jan 2001 (cvs revision 1.1.1.21 of xnu/iokit/KernelConfigTables.cpp). Since then, I get a kernel panic when I try to boot. This happened first when I let OSX auto-update the first time I installed the official release. It's still happening with the OSX 10.1 upgrade. Do I need to do something special? Should it be finding a module somewhere? I see AppleCore99PE seems to have broken off into a separate project... Thanks ---- Chris J. Bednar <http://optics.tamu.edu/~bednar/> Director, Distributed Computing Product Group http://AdvancedDataSolutions.com/

1 0

Re: devfs user client, RFC
by Godfrey van der Linden 15 Nov '01

15 Nov '01

G'day All, Just had an interesting idea. I have been bothered by the concept of a seperate server daemon to serve as a user land adjunct to a specfs (cdevsw) node. The problem is that I can't see anyway of giving user land code permission to write in an another tasks address space and THAT sort of power is very scary indeed. The alternative is to copy potentially vast amounts of data from client/kernel/server/kernel, expensive. The problem we are trying to solve is to get open/read/write/ioctl/close access to a /dev/... file. With the current implementation that means talking to a full on cdevsw entry, 'cause there is just no other alternative. This morning a thought occurred to me, I propose we change Libc syscall implemetation to redirect the read/write/ioctl/close calls selected by fd. Only the open goes directly into the kernel. Then we make the other cdevsw calls indirect through a pointer to a function in user land BEFORE taking the syscall trap. This solves all sorts of problems with getting data from process to process. The redirector will call functions that implement read/write/ioctl (RWI) on top of the CFPlugIn model proposed by ChrisS. Remember the the CFPlugIn interface to tapes and CD etc will already exist as that is MacOSXs preferred solution for device access of this nature. Am I going to die? Please be gentle with me in your responses. Godfrey

1 0

Is getpid supposed to return PID-1?
by Kevin Bohan 14 Nov '01

14 Nov '01

Hi, Thanks to everyone for the responses I got. Yes it was a bug in my code. I was simply calling getpid before the call to daemon(0,0). Once I moved the getpid further down it worked as expected. Kirby From: "Kevin Bohan" <kirby132(a)hotmail.com> To: darwin-kernel(a)lists.apple.com Subject: Is getpid supposed to return PID-1? Date: Tue, 13 Nov 2001 14:29:56 +0000 Hi, I have a BSD question and I apologise if this is not the right list, as this is not really my area. However I have a daemon that writes its PID to an external file when it starts up and I have a script that looks for this file, extracts the PID in order to provide a commandline interface that allows the process to be killed. However, when I call getpid to write the PID to the text file it always seems to write a value that is 1 less than the PID shown when I use ps to view the process in question. Also if I call kill on the value returned by getpid it tells me that the process doesn't exist. It would seem that ps is right an getpid is wrong. I can obviously just add one to the value returned by getpid, but is this safe/future-proof? Is this a known issue? What should I do? Kirby _________________________________________________________________ _________________________________________________________________ Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp

1 0