March 2009 - Darwin-dev - lists.apple.com

RE: Recommendations on uniquely identifying a system.
by Tom Fortmann 04 Mar '09

04 Mar '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Thread-index: AcmdF8+gSeV/ZPSRQ3KzDL3TVKAXXgAAXo6w That was my first thought, but we have a requirement that the value persist across installs, and I don't like leaving a preference file behind on uninstalls. So, using one or more of these values seems to be the best solutions. I'm just trying to find out if Apple is moving away from Mac addresses for this purpose. With no real documentation on the IOPlatformUUID value I think I will use a combination of the Mac address and serial number a unique ID string. I may factor in the IOPlatformUUID value if for some reason the serial number is missing. ___________________________________ Thomas Fortmann Director of Development Xcape Solutions, Inc. 207 Crystal Grove Blvd. Lutz, FL 33548 TEL 847 228 9831 www.xcapesolutions.net tfortmann(a)xcapesolutions.net -----Original Message----- From: Terry Lambert [mailto:tlambert@apple.com] Sent: Wednesday, March 04, 2009 4:23 PM To: Tom Fortmann Cc: darwin-dev(a)lists.apple.com Subject: Re: Recommendations on uniquely identifying a system. On Mar 4, 2009, at 12:39 PM, Tom Fortmann wrote: > I'm looking for recommendations on the best or most current way of > uniquely > identifying a system when registering with a server application. To > date > I've found three pieces of data that might do the job: > > 1. Primary MAC Address > 2. IOPlatformSerialNumber > 3. IOPlatformUUID value > > The MAC address and serial number are both well documented in the > technical > note TN1103. The platform UUID value I found using the > IORegistryExplorer > along with some references online regarding byhost preferences. Not > sure > what that is, but based on the article it sounds like this is the > preferred > value (over MAC address) for this purpose. Does anyone have more > insight on > this value and its intended use? The one thing that scares me is a > posting > I found for a problem with this value changing on every boot. > > As a little background I'm not looking to use this for licensing or > DRM. > We're just using it to register our agents with a back-end server that > collects status information. So, it needs to be fairly stable and > very > unique. If it changes with a hardware update or OS reload we are OK > with a > new entry. The old one will age out of the system. But we don't > want to > end up with multiple systems using the same value. If this is your criteria, then why not just generate your own UUID the first time you run and squirrel it away in a plist in your app somewhere, so it's relatively persistent over a given installation of the OS? -- Terry _______________________________________________ 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

Re: Recommendations on uniquely identifying a system.
by Terry Lambert 04 Mar '09

04 Mar '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com 1. Primary MAC Address 2. IOPlatformSerialNumber 3. IOPlatformUUID value -- Terry _______________________________________________ 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… On Mar 4, 2009, at 12:39 PM, Tom Fortmann wrote: I'm looking for recommendations on the best or most current way of uniquely identifying a system when registering with a server application. To date I've found three pieces of data that might do the job: The MAC address and serial number are both well documented in the technical note TN1103. The platform UUID value I found using the IORegistryExplorer along with some references online regarding byhost preferences. Not sure what that is, but based on the article it sounds like this is the preferred value (over MAC address) for this purpose. Does anyone have more insight on this value and its intended use? The one thing that scares me is a posting I found for a problem with this value changing on every boot. As a little background I'm not looking to use this for licensing or DRM. We're just using it to register our agents with a back-end server that collects status information. So, it needs to be fairly stable and very unique. If it changes with a hardware update or OS reload we are OK with a new entry. The old one will age out of the system. But we don't want to end up with multiple systems using the same value. If this is your criteria, then why not just generate your own UUID the first time you run and squirrel it away in a plist in your app somewhere, so it's relatively persistent over a given installation of the OS? This email sent to site_archiver(a)lists.apple.com

1 0

Samba, network interfaces, launchd, etc.
by David Farber 04 Mar '09

04 Mar '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com [global] ; interfaces = vmnet1 interfaces = 192.168.144.0/24 bind interfaces only = yes But this doesn't work on 10.5.4. Thanks. David _______________________________________________ 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'd like to make Samba listen on a specific network interface instead of listening on all interfaces. According to the Samba docs, I need to do something like this: $ sudo lsof -Pi | grep smbd smbd 1127 root 19u IPv4 0x5446e64 0t0 TCP *: 445 (LISTEN) smbd 1127 root 20u IPv4 0x681d66c 0t0 TCP *: 139 (LISTEN) Since Samba is under launchd control and the sockets it is listening on are passed to it, my first question is, should I and can I specify the interface I want in the smbd.plist? Looking through the launchd docs, I didn't see anything that looked like it would specify an interface. So maybe specifying the interface is the responsibility of the daemon. If so, then my second question would be if it would be appropriate to file a radar? (The interface I want to listen on is VMWare Fusion network interface that is bound only to the virtual machine that it is running. In other words, I want to use Samba to privately share files with a virtual machine that is running locally. I *don't* want Samba running on my local network.) This email sent to site_archiver(a)lists.apple.com

1 0

Re: Recommendations on uniquely identifying a system.
by Shawn Erickson 04 Mar '09

04 Mar '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:cc:content-type :content-transfer-encoding; bh=Hd+m/bdWqs6jRRX3IKGVDQ+Sl+5qnVZQ25S3n6H1aaA=; b=T08VFl1p4J6ELFVyy3uGbI6a/i81Wfern4AYOWnBf+LvPwAXvE1J90hhz0jwI/Zmz4 h3LH/gh6lm7oj2y59KChopdmGVllUyHanBh3QCz9GIhLO7Ke9PX6vWwAsn7P82zneTIj oOE4yIiSqci/ALbPsmFPgP6s9/2ks2mlTfDc4= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type:content-transfer-encoding; b=Tvj+3XicE29GeG8SR9AzmDeIpB3L8s2pLxTVYQ8SAP52wHX0k641kobnB06I84Kgxr OKb/rqrQztIdg6ATIKdkSAQE0B6pDRPiyXB4I8WtsIiL9kfMcC3tSzRNN6fYgqhfnzop Q2MCOBZ5+JnJuEO8cb2ln9po69ADvXT0TE22w= On Wed, Mar 4, 2009 at 12:39 PM, Tom Fortmann <tfortmann(a)xcapesolutions.net> wrote: > I'm looking for recommendations on the best or most current way of uniquely > identifying a system when registering with a server application. To date > I've found three pieces of data that might do the job: > > 1. Primary MAC Address > 2. IOPlatformSerialNumber > 3. IOPlatformUUID value > > The MAC address and serial number are both well documented in the technical > note TN1103. I would follow the recommendations of TN1103. The MAC address blended with platform serial number (or just one of those) should cover your needs well enough. -Shawn _______________________________________________ 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

Re: Recommendations on uniquely identifying a system.
by Eli Bach 04 Mar '09

04 Mar '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:received:received:message-id:from:to :in-reply-to:content-type:content-transfer-encoding:subject :mime-version:date:references:x-mailer; bh=EalAn5LuI2imEk+5Pc94Mkn5KaJjX3IeOl/PUTZ9z3k=; b=kjKOf0ZvnsuvgbsicpIzocFEJqdl3hUWy4Gv6DZZvRlxvnYW92XUZgF/1gVg0NqRQ7 hEcFqZeP+zgGFkdEnr3vXhNNBTYqlaoBpNIe9Yns3NxoO6fCgUJ2ZJBrMAU+UsweYML3 bKJa0sWuUhUwwjOr+qZ9OU3mw4+zC6PApOfvQ= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:from:to:in-reply-to:content-type :content-transfer-encoding:subject:mime-version:date:references :x-mailer; b=BK1XbzsafEz+4gGvgehqckQYRZ4l4A4f9Wtux+j0/uXQMk55x7dRjYM7ILYTIUF3hV ZEc2mXTM8yOS6OiB2BPzxXRpc4TmnoC+3P7UHundJs8L8pDV/sPCJQu7FetwcO9juBzZ lF2yFxDTp0T4ldQrkcCkfyD/zHkWedmuS1NIA= On Mar 4, 2009, at 12:39 PM, Tom Fortmann wrote: Eli _______________________________________________ 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… If it changes with a hardware update or OS reload we are OK with a new entry. The old one will age out of the system. But we don't want to end up with multiple systems using the same value. You can also dig down to get the serial number of the internal HD. But this depends on hardware configuration (do your computers have more than one internal HD, or perhaps internal SATA cards with other HDs attached... This email sent to site_archiver(a)lists.apple.com

1 0

Recommendations on uniquely identifying a system.
by Tom Fortmann 04 Mar '09

04 Mar '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Thread-index: AcmT97gA/61zSpXkSwGcuKl+4TxXTAJDOhfw I'm looking for recommendations on the best or most current way of uniquely identifying a system when registering with a server application. To date I've found three pieces of data that might do the job: 1. Primary MAC Address 2. IOPlatformSerialNumber 3. IOPlatformUUID value The MAC address and serial number are both well documented in the technical note TN1103. The platform UUID value I found using the IORegistryExplorer along with some references online regarding byhost preferences. Not sure what that is, but based on the article it sounds like this is the preferred value (over MAC address) for this purpose. Does anyone have more insight on this value and its intended use? The one thing that scares me is a posting I found for a problem with this value changing on every boot. As a little background I'm not looking to use this for licensing or DRM. We're just using it to register our agents with a back-end server that collects status information. So, it needs to be fairly stable and very unique. If it changes with a hardware update or OS reload we are OK with a new entry. The old one will age out of the system. But we don't want to end up with multiple systems using the same value. ___________________________________ Thomas Fortmann Director of Development Xcape Solutions, Inc. 207 Crystal Grove Blvd. Lutz, FL 33548 TEL 847 228 9831 www.xcapesolutions.net tfortmann(a)xcapesolutions.net _______________________________________________ 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

Re: GCC stack size
by Ariel Burton 04 Mar '09

04 Mar '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com User-agent: Mutt/1.4.1i > > I've been bitten by this problem in the past as well, especially in > functions that look a lot like the examples you've given (where an > outer if wraps several complex code blocks). The worst of it is that > in this example: > > void foo(int flag) > { > if (flag) { > largeStackUsingFunction(); > } else { > char buf[1000]; > /* ... other code which does not call functions ... */ > } > } What happens if you use alloca in the `else' in cases like this? Ariel Burton _______________________________________________ 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

Re: Information request on DWARF dsym reader
by Arnab Ganguly 04 Mar '09

04 Mar '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:cc:content-type; bh=B3K6macUhh2BZbcvTZzySsl1esVl7D/cQdA1YLBFMkY=; b=Jqf1iZJwsFKB99oab9A3u+3eaRYNVw3jrChe8VbEC3fMd1mi0KqqoF6XXmrjn0HL7r P2ZxOamq/FpmKrTcGgfm+Vmh3lHX50X0v0VYUHljLz9qJtWH2+YrLHQNBBC1FmIgPUPz C51rcqWvSQrN17g3Au1ROKkT91eVqCCz+vOrI= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; b=jgVBz7KFI7nLBbEDH89KRw9bLk1TpFdh4oKlhQxFQNNjcuIHsYcK9vtA9hAz6ZKNs1 7Oyg2fSe7tzLUg3Z0ggRlQqrmzm95vQMLA0iwB6WvI4RY31/jsxq2pEfxCiQ6lRH93+G vf9EgLmN/zHZZDU5HTEEDrJCoquKsWF6mMSOA= _______________________________________________ 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

Performing a task in the "middle" of a log-in and log-out process possible?
by Markus Hanauska 04 Mar '09

04 Mar '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com == Situation: == The Problems: == The Solution that would work: To describe that in more detail, the login process looks like that: The logout process looks like that: == The Question: -- Kindest regards, Markus _______________________________________________ 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… We have a user agent, that is started by launchd as soon as a user logs in. It then reads a preference file (~/Library/Preferences) and keeps the data cached in memory. While running, it constantly updates information in that file. The file is asynchronously written back to disk every now and then. When the user logs out, it receives a SIGTERM, updates the plist one last time, makes sure it's written to disk immediately and quits. So the preferences are read as soon as the process starts and they might change up to the very last second, till the process quits. This file should be kept in sync across multiple Macs. 1) To make sure the latest state of these preferences are backed-up, the back up must happen after this process has quit. Only then the pref file will really contain the latest changes. However every other user space process will be quit if the user logs out, first with a SIGTERM, then with a SIGKILL if it refuses to quit. 2) If this is a UI process, it can prevent the log out from taking place by refusing to quit, but then the agent process might never get killed in the first place and the pref file may not be up-to-date. 3) A root process running constantly in the background can notice if a user logs out, but it cannot necessarily access the file after log out. The user may have a FileVault, so access to the file after logout is impossible unless the process knows the mount password of the FileVault (insecure!). The user also may have a network home. Root processes cannot access data over networks in the first place. 4) The user may actually not log out, he may select to shutdown/ restart the system, in that case the root process also only would have a very limited time to sync the file. Since a sync can take arbitrarily long, it may get killed before it had a chance to really syn anything. 5) How can the pref file, that may have been back-up'ed (assuming problems 1 to 4 can be solved) to an external storage (e.g. an USB stick) be synced back to disk before the user logs in on the other Mac? If it is synced "after" the user logged in, the agent is already running, it has already read the pref file. Overwriting this file with new data will fail, as the changes won't be noticed by the already running agent. Issues 1 to 4 could be solved if there was a way to run a program similar like the compacting FileVault process. This process can run after the user has already logged out and it can delay a shutdown/ reboot till its operation completes... of course if the FileVault/ Network Home has already been unmounted, this won't work. Issue 5 could be solved if a process could be run after the user has logged in, but before any other user processes have been started. 1) The user logs in 2) If this is a FileVault or Network Home user, the home directory is mounted (either from the local FileVault DMG or via the network) 3) launchd starts the login agents for this user 4) Login Items of the user are started 5) The user gains control Here we'd need a step between 2 and 3. The user home is ready to be accessed by processes belonging to the user and a process is started that can delay step 3 till it is done with whatever operation it needs to perform. 1) The user logs out (or issues a shutdown/reboot) 2) All applications of the users are quit 3) All launchd agents of the user are quit/killed 4) If this is a FileVault or Network Home user, the home directory is unmounted 5) Either the login screen is started again or the system shuts down (possibly reboots after shutdown) Here we'd need a step between 3 and 4, so all processes (maybe except launchd) are already quit, the user has no control anymore, but we can still start a process belonging to this user to perform operations before step 4 and we can delay step 4 and 5 till it is done with whatever operation it needs to perform. Does anything like this exist? Can it be done without hacking the system to shrimps? And if this is not the right mailing list for asking such a question, what is the right mailing list? This email sent to site_archiver(a)lists.apple.com

1 0

Re: GCC stack size
by Ryan McGann 04 Mar '09

04 Mar '09

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Mar 3, 2009, at 2:00 PM, Brian Mastenbrook wrote: On Mar 3, 2009, at 3:35 PM, Ryan McGann wrote: Thanks, Ryan _______________________________________________ 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… We have some sample code that shows that the temporary returned from the function doMath is allocated stack space twice, when it can only be used once. But since the function in our actual production code is pretty complex, it's hard to know hwich problem we are experiencing. However, the above examples are both well documented cases of GCC problems; I actually got the basic ideas for both examples from Linus Torvald's (sometimes long) emails regarding Linux's problems with GCC stack allocation. I've been bitten by this problem in the past as well, especially in functions that look a lot like the examples you've given (where an outer if wraps several complex code blocks). The worst of it is that in this example: Yes, it looks like we are running into that problem. We don't have any large structures allocated on the stack, but a couple of ints here and there start to add up. In addition, it seems like GCC handles temporaries equally poorly. We have some code blocks that are very deeply nested due to switch statements and if/else blocks, and gcc seems to not know that certain temporaries it created for statements in the if condition can go away once the if statement has been evaluated. We have a lot of code that looks like: switch(something) { case blah: { if ((scanner->get_offset() + buffer_manager->get_offset() + scanner->start()) < cur_offset) { } } break; } GCC does all the math using temporaries, and then never realizes that the stack space for the temporaries can be re-used once the if statement has been evaluated. I tried turning off inlining--all those get() functions above are one line functions implemented in the header file, so gcc inlines them--but that seems to make little difference, which makes sense since all those "functions" are not really doing anything. There are still a few options left to you that I can think of. The first is to use LLVM - but in a slightly different way. Try running the code through llvm-gcc with the C backend, and fiddling with the different LLVM optimization settings. It's possible that llvm's better analysis of local variable liveness will result in (mangled and unreadable) C output that doesn't demonstrate the problem. Interesting. From the --help for llvm-gcc-4.2 I can't make much sense of it, or at least how to specify the output type. Whenever I tired it it seemed to just call gcc, but I'll have to play with it more. But since our library (and this function) is C++ not sure I can use this option anyway. Another option would be to hack up a transformer using clang (assuming your code is pure C), which would perform directed partial outlining on the functions you've identified to be problematic. Alas the code is C++. It's not *really* C++, as it doesn't even use virtual functions--it's more like structs with operator new() overriden to use lookaside lists, but still not pure C. This email sent to site_archiver(a)lists.apple.com

1 0