August 2007 - Darwin-dev - lists.apple.com

Re: Questions about launchd
by Josh Wisenbaker 24 Aug '07

24 Aug '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Aug 24, 2007, at 7:47 AM, Dave Zarzycki wrote: What needs to be done is this: 1) Create the file somewhere else 2) Once the file is in a sane state, move it to the queue directory Josh -- Josh Wisenbaker, ACSA http://www.afp548.com Breaking my server to save yours. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… ...and second thing. I used launchd also for moving some files to another computer. You are using path called "Queue Directories". That means if anything is placed into this directory (not sure about subdirectories, please confirm), the script is launched. Unfortunately (I am not sure if it is also launchd "feature" or not) if I am placing some big file (like pictures in print resolution, but in this case just imagine DVD ISO image), the launchd (or Lingon) isn't checking if the file is open or not. In my opinion it just checking HFS+ header and if some file is added (means record for that file is created - but not content) the script is launched which case in my example "unable to copy because too much characters" error. I did it in my script with checking "lsof" if the file is used or not. Isn't there any other way directly in launchd or Lingon ? Unfortunately, there are no facilities in the OS for knowing whether a file is open or not. I've also just had launchd run a script that checks the file for usage and then moves it... This email sent to site_archiver(a)lists.apple.com

1 0

Re: Questions about launchd
by Dave Zarzycki 24 Aug '07

24 Aug '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Aug 22, 2007, at 6:38 PM, Michal Taurich wrote: Hello, Best Regards Michal Hello Peter, What needs to be done is this: 1) Create the file somewhere else 2) Once the file is in a sane state, move it to the queue directory Good luck, davez _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… my name is Michal and I got contact to you from Peter Borg, creator of Lingon (launchd GUI interface). I am forwarding you my e-mail to him. It is not so much about Lingon it is basicaly about launchd. Any help in my cases will be appriciated. Thanks in advance. first of all I would like to thank you for creating such amazing tool like Lingon. It helps me very much. But I would like to ask you about one thing I clearly not understand. I am using for monitoring software called RMA ( RMA 1.23 for Mac OS X 10.4 (i686) - http://www.hostmonitor.biz/download/rma123_darwin88i32.zip) This is just an agent for Mac OS X. I created launchd plist with your Lingon. I tried to create "Keep an application/script always running" but when I look into Console it creates some error (exactly "launchd: RMA Host Monitor Script: 9 more failures without living at least 60 seconds will cause job removal ... will restart in 10 seconds). Unfortunately I am not able to find/log any significant error anywhere (Console, sterr output even if I enable Debug option). So where can I find exactly why the script / command failed? I also understand this is not problem of Lingon but mostly launchd itself. I successfully create the plist (and keep it running) only with "run a job at startup". But maybe you understand I want to be sure the RMA agent is running and if there will be any unexpected failure the plist is strictly written only "at startup" so that means during the cycle, the launchd didn't check if the script is running or not. So is there any way how to monitor if the script is running other than just look into launchctl and see it in list? Because I had several cases when script was listed in launchctl but was not active (means script was not running).Or must I create another bash script which will check if the process is running ? The RMA Agent is started with following command line: ./rma -d ~/RMA/rma.ini (-d is daemon, rma.ini is configuration file) It sounds like the fundamental problem is that the script or something that the script runs is daemonizing itself. Try this. Run the script by hand. Does it give you a shell prompt after a brief amount of time, or do you need to control-C the script to get your shell prompt back? ...and second thing. I used launchd also for moving some files to another computer. You are using path called "Queue Directories". That means if anything is placed into this directory (not sure about subdirectories, please confirm), the script is launched. Unfortunately (I am not sure if it is also launchd "feature" or not) if I am placing some big file (like pictures in print resolution, but in this case just imagine DVD ISO image), the launchd (or Lingon) isn't checking if the file is open or not. In my opinion it just checking HFS+ header and if some file is added (means record for that file is created - but not content) the script is launched which case in my example "unable to copy because too much characters" error. I did it in my script with checking "lsof" if the file is used or not. Isn't there any other way directly in launchd or Lingon ? Unfortunately, there are no facilities in the OS for knowing whether a file is open or not. This email sent to site_archiver(a)lists.apple.com

1 0

Questions about launchd
by Michal Taurich 23 Aug '07

23 Aug '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hello, my name is Michal and I got contact to you from Peter Borg, creator of Lingon (launchd GUI interface). I am forwarding you my e-mail to him. It is not so much about Lingon it is basicaly about launchd. Any help in my cases will be appriciated. Thanks in advance. Best Regards Michal Hello Peter, first of all I would like to thank you for creating such amazing tool like Lingon. It helps me very much. But I would like to ask you about one thing I clearly not understand. I am using for monitoring software called RMA ( RMA 1.23 for Mac OS X 10.4 (i686) - http://www.hostmonitor.biz/download/rma123_darwin88i32.zip) This is just an agent for Mac OS X. I created launchd plist with your Lingon. I tried to create "Keep an application/script always running" but when I look into Console it creates some error (exactly "launchd: RMA Host Monitor Script: 9 more failures without living at least 60 seconds will cause job removal ... will restart in 10 seconds). Unfortunately I am not able to find/log any significant error anywhere (Console, sterr output even if I enable Debug option). So where can I find exactly why the script / command failed? I also understand this is not problem of Lingon but mostly launchd itself. I successfully create the plist (and keep it running) only with "run a job at startup". But maybe you understand I want to be sure the RMA agent is running and if there will be any unexpected failure the plist is strictly written only "at startup" so that means during the cycle, the launchd didn't check if the script is running or not. So is there any way how to monitor if the script is running other than just look into launchctl and see it in list? Because I had several cases when script was listed in launchctl but was not active (means script was not running).Or must I create another bash script which will check if the process is running ? The RMA Agent is started with following command line: ./rma -d ~/RMA/rma.ini (-d is daemon, rma.ini is configuration file) ...and second thing. I used launchd also for moving some files to another computer. You are using path called "Queue Directories". That means if anything is placed into this directory (not sure about subdirectories, please confirm), the script is launched. Unfortunately (I am not sure if it is also launchd "feature" or not) if I am placing some big file (like pictures in print resolution, but in this case just imagine DVD ISO image), the launchd (or Lingon) isn't checking if the file is open or not. In my opinion it just checking HFS+ header and if some file is added (means record for that file is created - but not content) the script is launched which case in my example "unable to copy because too much characters" error. I did it in my script with checking "lsof" if the file is used or not. Isn't there any other way directly in launchd or Lingon ? _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Wait-Queue Panic with system level calls.
by JanakiRam 22 Aug '07

22 Aug '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Dkim-signature: a=rsa-sha1; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:mime-version:content-type; b=bE7A5mkNihEJOIawj/ArcmNNhD4tRqeAnOxhlypz0NiTDr779iQExPs6CL7oqB8qklESY7QEBMDr3VdKJ8aVwytCZb6/3LwzhbgIsqu7JfLjQZbTtxqY7e8uArCuM/s3fRS/20k2v/YnH3Hc1aJkFibSykfQ/15AGTYc9MwU1Jk= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:mime-version:content-type; b=tsxXI3jqZY3tsontQDEXozRh4HeqwWf+/x0ZmmuPTT6LXxzaPZ5UFVVtxrSkLqDYCsKZwhjEwXIpBrl7LHnuCQCVgBIuU09T7T2Ta4QLZP6QolT6rajn26pAhd0udgVN4kgjLenjcy9u+DczZfmPKBXdxUrZBW9wbxWqb4crpXM= _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Mac OS X System (10.3.9) deletes itself - bug or virus?
by Swen Metzger 22 Aug '07

22 Aug '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Thanks in advance Swen _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… I encounter a weird system behavior; the system (10.3.9 Build 7W98, Powerbook G4 17") deletes itself. Now for the fourth time (since a month or so). If the computer is on, it always starts at 3AM when the cron job starts, compressing all log files etc. But accidentally(?) starts cleaning up the whole system partition (everything below the main directories in /. I first thought I made somehow a misstake (e.g. creating a soft link in / which shouldn't be there). The second time I was not sure, too (I found a strange .file which shouldn't be there and which was new and not created by myself). But now, after the fourth time, I suspect something else (there is again a file I am not sure about; /.hotfiles.btree with root admin 65536; if I remember right with different name than before). For security reasons, I always install the system on a separate partition. In this case, it prevented that all my files were deleted. Only insensitive system files are affected, which I can easily reinstall by booting from an external firewire disk (used as backup and always with me). Nevertheless, it is annoying! Could it be that somehow a re-installation of the system leads to this odd behavior? Two things I recently did; a software update via Internet (Option Apple -> Software Update), and, I used the restore option of disk utility for the backup (and re-installion) of the system after I changed to a 160GB disk (new and empty, internal). Since then all the trouble ... And, also since then, for some reasons, the option 'More Info ...' (Apple -> About this Mac') is disabled, too. Any suggestion to fix this would be highly appreciated (before upgrading to Leopard / new MacBook :) ... This email sent to site_archiver(a)lists.apple.com

1 0

Re: "Bad CPU type" when building on G4
by John Zorko 21 Aug '07

21 Aug '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Terry, Thanks for the response! Regards, John Falling You - exploring the beauty of voice and sound http://www.fallingyou.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… You are probably building both Intel and PPC as 64 bit, and you are lucking out on the Core 2 Duo being a 64 bit CPU and not noticing. Normally you have to specifically ask for 64 bit via -arch ppc64 - arch x86_64; if you are doing this, then you probably also want - arch i386 -arch plc ppc to build it 4-way universal (if you need it to be 64 bit for the address space on 64 bit machines, or to get twice as many registers for the compiler to use on Intel for performance reasons, etc,.). Actually, I fixed it just by specifying -mcpu=G4, but I suspect that this is a lowest-common-denominator approach (at least across PPC -- I was surprised to see that -mcpu=G4 didn't have any adverse effect on the Intel Core2 Duo build, though I also suspect that I might have the same issues if I tried to build it on a Core Duo / Solo). I'll try adding all of those arches, and see if I can come up with one binary image that works on G4 / G4 / Core / Core2 :-) This email sent to site_archiver(a)lists.apple.com

1 0

Re: "Bad CPU type" when building on G4
by Terry Lambert 21 Aug '07

21 Aug '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com -- Terry On Aug 20, 2007, at 3:28 PM, John Zorko <jmzorko(a)mac.com> wrote: Hello, all ... Regards, John Falling You - exploring the beauty of voice and sound http://www.fallingyou.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/tlambert%40apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… You are probably building both Intel and PPC as 64 bit, and you are lucking out on the Core 2 Duo being a 64 bit CPU and not noticing. Normally you have to specifically ask for 64 bit via -arch ppc64 -arch x86_64; if you are doing this, then you probably also want -arch i386 - arch plc ppc to build it 4-way universal (if you need it to be 64 bit for the address space on 64 bit machines, or to get twice as many registers for the compiler to use on Intel for performance reasons, etc,.). I'm working on an application that runs on Intel / PPC Macs (OSX) as well as Windows. When I build this app on my MacBook (Core2 Duo) or G5 (2.3ghz 970MP at work, dual 2.7ghz at home), the app runs. However, when I try to build and run it on a G4, it builds, but does not run. The error returned is "Bad CPU type for executable." The issue seems to be 64-bit vs. 32-bit. I've googled around and found references to --disable-multilib, but these references were always with respect to bootstrapping gcc. I'm not trying to build a new gcc, I just want to use the one that comes with OSX (Apple's gcc4.0.1). Will just passing "-m32" in the gcc parameters fix this? I'm trying this now, but this app takes hours to build on a G4 (1.33ghz 12" PowerBook), and I figure I might get an answer here before it's finished :-) This email sent to tlambert(a)apple.com This email sent to site_archiver(a)lists.apple.com

1 0

"Bad CPU type" when building on G4
by John Zorko 20 Aug '07

20 Aug '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hello, all ... Regards, John Falling You - exploring the beauty of voice and sound http://www.fallingyou.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… I'm working on an application that runs on Intel / PPC Macs (OSX) as well as Windows. When I build this app on my MacBook (Core2 Duo) or G5 (2.3ghz 970MP at work, dual 2.7ghz at home), the app runs. However, when I try to build and run it on a G4, it builds, but does not run. The error returned is "Bad CPU type for executable." The issue seems to be 64-bit vs. 32-bit. I've googled around and found references to --disable-multilib, but these references were always with respect to bootstrapping gcc. I'm not trying to build a new gcc, I just want to use the one that comes with OSX (Apple's gcc4.0.1). Will just passing "-m32" in the gcc parameters fix this? I'm trying this now, but this app takes hours to build on a G4 (1.33ghz 12" PowerBook), and I figure I might get an answer here before it's finished :-) This email sent to site_archiver(a)lists.apple.com

1 0

Re: IOKit KEXT Questions
by Andre-John Mas 18 Aug '07

18 Aug '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hi, Andre On 15-Aug-07, at 18:21 , Matt Burnett wrote: I have 2 questions regarding IOKit KEXTs. Example HelloIOKit.cpp Code: ... extern "C" int execve(void *, void *, int *); ... IOLog("Found execve at %p\n",execve); ... Example kextload output: kextload: extension HelloIOKit.kext appears to be valid kld(): Undefined symbols: _execve ... Thanks -Matt Burnett _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/ajmas%40sympatico.ca _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… As stated previously giving us an idea of what you are trying to achieve would be the best solution to getting a good answer. I realise that you may be trying to create a solution that you don't want to publicise too soon, but if there is any way you can explain your goal, even if it is part of the solution I think it would be the best way to go. In many solutions certain decisions have been made for better or worse. Generally the approach is accepted unless there is a good reason to bring them into question. With an API certain compromises are made in trying to allow maintainability and providing developers with a stable interface. I don't think anyone is accusing you of anything, though at the same time since we don't understand your goal we can't help you until we understand. There are some things that can't be answered properly unless the context of the question is provided. I myself have asked questions that were obvious as what I was asking, until I tried seeing from the other person's point of view. - How does IOMemoryDescriptor deal with memory protection (mprotect/ vm_protect)? Is there a class available to check/change the protection of certain pages in memory or should i use mprotect/ vm_protect? - Why do IOKit KEXTs have issues resolving kernel symbols, and what can i do to work around it. For example if i create a KEXT in C, i can lookup the address of any symbol which is in /mach.sym, however I get unresolved symbol errors if i do the same in a IOKit KEXT. This email sent to ajmas(a)sympatico.ca This email sent to site_archiver(a)lists.apple.com

1 0

Re: IOKit KEXT Questions
by Ron Aldrich 16 Aug '07

16 Aug '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Aug 16, 2007, at 1:46 PM, Ernesto Corvi wrote: Ron _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… I actually want the opposite. I want the the kernel not to be generically patchable, as it is right now, and be able to have a KPI, such as kauth (although kauth is severely limited in scope right now) where benevolent users can actually implement the kernel hooks they need. >From a user standpoint, I want kernel development to be as restricted, and well controlled as possible. I know that as a kernel developer, it is extremely difficult to add functionality to the kernel without damaging it's stability. This whole discussion reminds me of the bad old days of System 7, where developers considered anything fair game - just look at old versions of Microsoft Office for an example of how badly kernel (or system) patches can go. This email sent to site_archiver(a)lists.apple.com

1 0