Lap Cat Software Blog

I apologize for not posting this earlier. I’ve been extremely busy lately, and I had discussed the issue with someone who said that he or she (he) was going to post about it (but hasn’t).

The security alert is for FileVault users running Mac OS X 10.5.2. You thought that FileVault encrypted your personal data, right? Wrong! In Mac OS X 10.5.2, the location of the CFNetwork caches was moved from ~/Library/Caches, which is within your home directory and thus encrypted, to /private/var/folders, which is not within your home directory and thus not encrypted. This means that anyone with physical access to your hard drive could, for example, determine which URLs you’ve loaded, even if your computer is shut down.

Note that nothing about this change was mentioned in the Mac OS X 10.5.2 release notes.

For further reference on this issue, see the thread that began in the WebKit SDK mailing list and was moved by me to the Macintosh Network Programming mailing list. Thanks to Eric Long, who noticed the change in the first place, and Ron Hunsinger, who performed testing that I was too lazy to do. (In my defense, I haven’t yet migrated my FileVault account from Tiger to Leopard, so the issue doesn’t affect me directly.)

This post is dedicated to E. Gary Gygax, the second greatest corrupter of youth in history. It’s about D&D, that is, DWARF and dSYM. As of 2008-02-27, the STABS debugging symbols format has been deprecated by Apple. The default value for the DEBUG_INFORMATION_FORMAT build setting in Xcode projects had been stabs, but now it’s time to move on. (I’m talking to you, Justin Long.) Our other options are dwarf or dwarf-with-dsym. Also cake or death.

With STABS, you could build the release version of your app with debugging symbols, make a copy of the executable MyApp.app/Contents/MacOS/MyApp to keep, strip the executable for shipping, and then use the unstripped executable for symbolizing crash reports by giving a space-separated list of stack trace addresses to the command-line tool atos. Unfortunately, atos cannot currently serve this purpose with DWARF. Unlike STABS, DWARF does not include the debugging symbols in the executable itself but merely includes references to the intermediate object files, which do contain debugging symbols. You can usually find these .o files in a sub-directory of the build/MyApp.build directory. If you delete the object files after building with dwarf, you won’t be able to step through your app’s code. (With stabs, the object files are refuse.) You also won’t be able to step through the code if you strip debugging symbols from your app, even if you keep the object files, because the references to the object files will be gone from the executable.

To avoid losing the debugging symbols for your app after stripping, you want to use the option dwarf-with-dsym. The DWARF with dSYM option performs an additional step beyond ordinary DWARF: it creates a separate MyApp.app.dSYM file that contains all of the debugging symbols for your app. In fact, the DWARF with dSYM option allows you to step through your code regardless of whether the executable is stripped! This is possible because gdb will look for the .dSYM file in the same directory as your app. It doesn’t need to know the name or location of the object files. If you don’t strip debugging symbols, you can use either the .o files or the .dSYM file for debugging, but for the local debug build of your app there’s no point in using dSYM, since that would just prolong your build time. You have better things to do than wait for builds, such as writing comments on Slashdot.

The trouble with atos is that it does not reliably find debugging information in .dSYM files for stripped executables. Although Apple’s documentation (as of 2007-04-02) says, “If you’re using DWARF dSYM files, you must be using the version of atos included in Xcode 3 (Mac OS X version 10.5)”, Apple’s engineers say, “The underlying framework that atos uses doesn’t support loading symbol names from dSYM files in Leopard.” In my testing, however, there doesn’t seem to be a difference between Leopard and Tiger, at least not with Xcode 2.5 on Tiger. On both Leopard and Tiger, atos successfully loads symbol names from .dSYM files (I deleted the .o files) for unstripped executables. For stripped executables, in contrast, atos frequently fails to load the symbol names, or even gives inaccurate results.

The CrashReporter Technical Note suggests loading your app and its .dSYM in gdb to translate stack trace addresses from crash reports. That’s like having to start your car in order to read the odometer. (Oh wait, I have to do that, Nissan!) An alternative method is the command-line tool dwarfdump. It requires only the .dSYM file, not a copy of your app, and its --lookup option will do the same job as gdb without the overhead.

Please note that by breathing, blinking, or moving at all, even to command-w this page, you thereby register your agreement not to disclose or discuss this information anywhere with anyone at any time, no matter the duress, torture, or water-boarding you may undergo to extract it. This agreement holds despite the fact that the information is publicly available on the internet for every person in the world to read. Failure to uphold this agreement will result in multiple, painful cat scratches, in certain cases leading to cat scratch fever.

At the risk of provoking their ire and being branded a moron, I wish to dispute a claim that has been made several times by Rixstep about a supposed security vulnerability in Leopard’s NSDocumentController: Cocoa’s document controller overrides file system permissions without authentication. As far as I can tell, this claim is false.

I should note at the outset that I have nothing personal against Rixstep. Their syndicated feed has long been among my (numerous) favorites that I subscribe to in Vienna, as you can see by downloading the exported opml from my blog. I welcome legitimate criticism of Apple, and I have found Rixstep’s articles entertaining in the past, though of late they have become overly juvenile. As far as the whole ‘Cross-Platform Bait & Switch’ incident is concerned, I can’t comment on the legality of reproducing the quotations, because I’m not a lawyer, but I don’t think that Rixstep can be accused of misquoting or quoting out of context in this case.

Anyway, I’ve included the full text below of the email I sent to Rixstep a month ago. Retraction was perhaps too much to hope for, but I thought they would at least stop making the false claim after reading my email. They haven’t, which is why I’m now ‘going public’. (This should knock Britney off the front page.)

Hi. I read your article at <http://rixstep.com/2/1/20071227,00.shtml>, and it got me a little worried, so I did some testing. My results are that NSDocumentController in Leopard does not allow you to override Unix permissions. In fact, NSDocumentController in Leopard is more strict than the Unix permissions: it won’t allow you to save a writable document when you don’t have write permissions for the enclosing directory.

You are correct that saving a document always deletes the existing file and creates a new one with a different inode, and you are also correct that some of the user-visible NSDocumentController warning messages are misleading. However, if you don’t have write permissions for a directory, then NSDocumentController won’t let you delete or add a file, and even if you do have write permissions for the directory, NSDocumentController won’t let you delete someone else’s file or add a file with the same name if the sticky bit is set.

Thus, I believe that the ramifications for system administrators are negligible. With certain Unix permissions, it has always been possible to delete a file in a directory and replace it with a new file of the same name. Leopard has not changed that at all, so system administrators should take the same precautions against this scenario that they have always taken.

-Jeff

You don’t have to take my conclusions for granted, though. In a matter of minutes, you should be able to throw together a bare-bones document-based application suitable for testing the behavior yourself.

I have no desire to discourage criticism of Apple or Leopard. There are major problems in Leopard, and I don’t yet find it acceptable for use as my primary operating system. I just think that this spurious security issue obscures the real issue of whether the new Leopard NSDocumentController behavior is desirable.

P.S. If you really want to talk about a waste of precious disk space, /Applications/Mail.app/Contents/Resources is an astounding 277 MB on Leopard.

The release of Leopard has given third-party developers a lot to do: attempting to restore features lost from Tiger, for instance. (By the way, where is the second party, and why am I never invited?) My friend Rainer Brockerhoff has provided a way, or Quay, to display hierarchical popup menus in the Dock again. One of my most missed features in Leopard is using NSLog to spew output exclusively to Xcode’s console log. When you debug or run your app in Xcode on Tiger, you can put NSLog calls everywhere without worrying about polluting console.log. In my opinion, console.log is only for important messages and errors. I frequently ask users to consult it if they’re experiencing a problem with an app. Either that or the Oracle at Delphi.

Leopard dispenses completely with console.log, though there is a “Console Messages” database query in Console. Whereas on Tiger stdout and stderr standardly go to console.log, on Leopard they boldly go to system.log (as well as to the “Console Messages” query). On either version of Mac OS X, Xcode redirects stdout and stderr to its own console log, so they don’t appear in Console at all.

According to the documentation, NSLog sends a message to stderr. This is true for Tiger, and it’s also true for Leopard, but Leopard’s NSLog has the additional behavior of sending a message to system.log regardless of whether stderr is redirected. Thus, when you debug or run your app in Xcode (these may amount to the same thing in Xcode 3), messages from NSLog appear both in Xcode’s console log and in system.log! Curiously, there is no duplication of NSLog messages in system.log when stderr is not redirected.

If you prefer to keep your debug output out of system.log, the workaround for this new NSLog behavior is to abandon NSLog for debugging purposes on Leopard. After much experimentation with asl, I realized that our old faithful printf would work. Since printf writes to stdout, its output is redirected by Xcode. Plus, when you’re debugging your app in Xcode you don’t really need NSLog to tell you the name of your app, the date, or your shoe size.

A limitation of printf is that it doesn’t handle the format specifier %@ for an Objective-C object. With Cocoa, therefore, we want an Objective-C wrapper around printf (like, um, NSLog). If you add the following code to your target’s .pch file, you’ll have an Objective-C debug logging function JJLog available throughout your target’s code. To enable logging in your app’s debug build, just add JJLOGGING to the GCC_PREPROCESSOR_DEFINITIONS setting (AKA “Preprocessor Macros”) in the debug build configuration.

#ifdef __OBJC__
	#import <Cocoa/Cocoa.h>
	#ifdef JJLOGGING
		#define JJLog(...) (void)printf("%s:%i %s: %s\n", __FILE__, __LINE__, __PRETTY_FUNCTION__, [[NSString stringWithFormat:__VA_ARGS__] UTF8String])
	#else
		#define JJLog(...)
	#endif
#endif

In your app’s release build, the debug function is a NOP that the compiler will almost certainly optimize out. This conditional code should not cause problems when using GCC_PRECOMPILE_PREFIX_HEADER, because Xcode already generates a separate precompiled prefix header for each build configuration. See the .pch.gch.hash-criteria files in /Library/Caches/com.apple.Xcode.###/SharedPrecompiledHeaders.

You can send gobs of gab to JJLog without repercussion or remorse. However, you’ll still want to use NSLog (sparingly, please) for runtime errors in your release build. Now to continue in the spirit of this post, I’ll redirect the epilogue to /dev/null.

Some people are apparently under the impression that firewall rules are the be-all and end-all of Mac OS X security. Deny all from any to me in. Before you spend days composing the perfect ipfw ruleset, however, take a look at sysctl. (You could spend days looking at sysctl.) One of the biggest misconceptions about Mac OS X — a misconception encouraged by the System Preferences UI — is that the firewall is not enabled out of the box. Nonsense!

net.inet.ip.fw.enable: 1

Admittedly, it’s a fairly flammable firewall (allow ip from any to any), but it’s enabled. More important, the default sysctl settings do provide at least a modest level of protection. For example:

net.inet.icmp.timestamp: 0
net.inet.icmp.maskrepl: 0

These prevent your machine from responding to ping -M time and ping -M mask requests. What time is it? Where the party at?

By Mac OS X herein I shall mean Leopard and Tiger. (Panther is not yet abandonware, but at this point it merely receives Christmas cards.) When you ’start’ the firewall in System Preferences, Tiger adds a number of rules to ipfw. Leopard, on the other hand, includes a new Apple-designed firewall that mostly replaces ipfw. Though Leopard respects custom ipfw rulesets, it adds no rules to ipfw when you choose to block incoming connections. I don’t know much about how the Leopard firewall works; presumably it protects you somehow or other.

Leopard does still employ ipfw for “Stealth Mode”. If you turn that on, both Leopard and Tiger add the rule deny icmp from any to me in icmptypes 8. Curiously, Stealth mode prevents you from ‘pinging yourself’ in Leopard but not in Tiger. For fun (according to a rather loose definition of fun), try ping localhost. Somebody bring me a mirror! The difference in behavior is due to the Tiger-only rule allow ip from any to any via lo* that precedes all other rules.

Looking at the ipfw rules, you might think that Stealth Mode is overrated, like Notre Dame football and most wines over three dollars. If you want to learn its true value, you need to consult sysctl.

net.inet.tcp.blackhole: 2
net.inet.udp.blackhole: 1

With these settings, your computer mercilessly crushes attackers into singularities, as well as dropping their packets without response when sent to closed ports.

It would be nice if Bonjour wasn’t so talkative, broadcasting its greetings to your entire LAN. Bonjour also complains loudly when told to shut up by the firewall. On the other hand, you’re going to have to broadcast anyway for DHCP. Moreover, your LAN will see your internet traffic, especially on wireless.

Whoa, I suddenly realized after all these years that if you’re a grown man who lives in an attic, hangs out with high school students, has an ‘office’ in the bathroom, and is known for one syllable utterances, you’re not cool. You’re a loser.

Although I abhor self-promotion — much as Roger Federer abhors winning tournaments — someone must take on this thankless task. (I had to fire my publicist, because he had never heard of me.) Thus it is with great regret and sorrow that I announce my latest contribution to the WordPress open source project. It was just one minuscule twitch for mankind, yet one ginormous vault for a man, viz., yours truly.

Now some critics might claim that the security issue was trivial. In my defense, I would argue that critics are doo-doo heads. Except the ones in New York: they all love me.

Although it seems late in the game to be adding new features to Tiger, Mac OS X 10.4.11 includes not only the high-profile Safari 3 update but also a few other interesting little additions. For instance, when you uncompress a .tar.gz file in Finder it no longer leaves behind a useless .tar that you have to manually trash. Neat!

Of course, it wouldn’t be Finder without some annoyance. The trade-off for un-cluttering your Desktop is up-cluttering your console.log:

BOMArchiveHelper[454] opened /Users/jeff/Desktop/.BAHEm7Xn/my-directory.tar

I’m not sure why they didn’t use tar xzf, which is approximately a billion times faster. Baby steps…