December 2005 - Darwin-dev - lists.apple.com

Catching and preventing hard sleep
by apple_lists＠mobileneurons.com 02 Dec '05

02 Dec '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Greetings, // Include <IOKit/pwr_mgt/RootDomain.h> // Prevents hard sleep when closing lid. found in Insomnia::init() IOPMrootDomain *root = getPMRootDomain(); root->receivePowerNotification(kIOPMDisableClamshell); // Allows hard sleep when closing lid. found in Insomnia::free() IOPMrootDomain *root = getPMRootDomain(); root->receivePowerNotification(kIOPMEnableClamshell); * http://binaervarianz.de/projekte/programmieren/meltmac/ I appreciate any and all commentary. Most humble regards, Eric Coffin mobileneurons _______________________________________________ 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 have been reading through documentation and mailing list archives for sometime in search of a particular bit of functionality specific to apple laptops. I need to catch and veto hard sleep caused by the closing of a laptop lid on system running OS 10.4. From my research I understand that this may be achieved through the Kernel, likely through a KEXT. I have looked at the example RegisterForSleep, which shows how a program can register to listen for and handle system sleep calls. This allows an application to prevent idle sleep and gives them the oppourtunity to do some housekeeping before an impending hard sleep, caused by closing the laptop, or clicking sleep from the apple menu. I have also looked at the KEXT called Insomnia* which when installed prevents the system from going into hard sleep when the laptop lid is closed. This is achieved by communicating with the root power domain: Though preventing sleep is part of the solution I need to implement, more importantly I must be able to catch it, prevent it. Attempting a combination of RegisterForSleep and Insomnia did not do it. I loaded the KEXT when the event kIOMessageSystemWillSleep was sent to my program but to no avail. The system will sleep. I have read a few brief snippets within the mailing list that said it is possible to catch a hard sleep at the kernel level and veto it, but I have yet to discover how to do this. In an older version of <IOKIT/pwr_mgt/IOPM.h> I came across the following enumeration definition: // Power events enum { kClamshellClosedEventMask = (1<<0), // User closed lid kDockingBarEventMask = (1<<1), // OBSOLETE kACPlugEventMask = (1<<2), // User plugged or unplugged adapter kFrontPanelButtonEventMask = (1<<3), // User hit the front panel button kBatteryStatusEventMask = (1<<4) // Battery status has changed }; kClamshellClosedEventMask looks very interesting. I don't have a sense of how I could setup a event listener for this though. As well, the current IOPM.h file on my system is actually without this definition. This email sent to site_archiver(a)lists.apple.com

1 0

Notification handling without run loop of a primary thread
by Vladimir Magaziy 02 Dec '05

02 Dec '05

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com Hi all, I am trying to handle distributed notifications in a CoreFoundation Tool that does not run a run loop of its primary thread (actually, it is the requirement). I can not find any solution that works, so I ask you about possibility to do in general. I thought that it is possible to track notifications using a run loop of a separate thread (see code snippet below), but CFRunLoopRun function does not run a thread run loop, it simply returns a control and a thread finishes. I guess that I do not initialize a thread run loop correctly, so please suggest me the appropriate way to run it and force to handle notifications. /*******************************************************/ #include <CoreFoundation/CoreFoundation.h> #include <pthread.h> int gFlag = 0; void gNotificationHandler(CFNotificationCenterRef center, void *observer, CFStringRef name, const void *object, CFDictionaryRef userInfo) { gFlag = 1; } void *gThreadRoutine(void *value) { CFNotificationCenterRef center = CFNotificationCenterGetDistributedCenter(); CFNotificationCenterAddObserver(center, NULL, gNotificationHandler, CFSTR("NAME1"), CFSTR("NAME2"), CFNotificationSuspensionBehaviorDeliverImmediately); CFRunLoopRun(); CFShow(CFSTR("[THREAD] FINISHED")); return NULL; } int main() { pthread_t thread = {0}; pthread_create(&thread, NULL, gThreadRoutine, NULL); while (1 != gInt) {/* do nothing */ } return 0; } /*******************************************************/ Thanks in advance! Best regards, Vladimir Magaziy. _______________________________________________ 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: weird IOKit crash in IOIteratorNext()
by Julian Mayer 02 Dec '05

02 Dec '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com _IOIteratorNext: +00 mfspr r0,lr +04 stw r0,0x8(r1) +08 stwu r1,0xffa0(r1) +0C addi r4,r1,0x40 +10 bl 0x905cd014 ; symbol stub for: _io_iterator_next +14 cmpwi cr7,r3,0x0 - its prebound, incorrectly. - its been stomped by a memory smasher. My bet, from having witnessed this happen before, is the first. -Ed If you look at the disassembled code for IOIteratorNext in IOKit.framework, you'll see there is no way this call sequence should happen. IOIteratorNext does not call CFRelease at + 0x10, it calls _io_iterator_next by way of a dyld_stub, which loads a function pointer and jumps to it. There are only two reasonable ways that dyld's function pointer could be wrong. hello ed thanks for your analysis! it is still not clear to me why the binary in question crashes on one machine and runs properly on all other...is there a real problem and the prebinding problem only accounts for an incorrect crash log, or is prebinding completely responsible for the problem? if so, what do i have to do to remedy this problem? does this also mean that the rest of the crash log is also wrong? i've not taken any special steps in regards to prebinding, i've just used Xcode's defaults for the deployment mode, which i thought are fine...? thanks, julian _______________________________________________ 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: Determining the canonical case for a file name on HFS+
by Rehan Rizvi 02 Dec '05

02 Dec '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Thread-index: AcX225zv9FDsZ8lYQfS0tGQTmMXXHQ== Thread-topic: Determining the canonical case for a file name on HFS+ Hello Camillo, You wrote: "I need to know this for the purpose of comparison, and I cannot simply do a case-insensitive comparison because I do not know if / folder/ is on a case-insensitive or case-sensitive volume." I was just wondering, what is stopping you from finding out what kind of volume it is? The information is there in the volume header. If you want to find out more about volume headers you can do a search on the developer.apple.com for "tn 1150" or HFSPlusVolumeHeader. Thanks, Camillo _______________________________________________ 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: flushing HFS cache to disk (Dave Yost)
by Dominic Giampaolo 02 Dec '05

02 Dec '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com [ sync() being synchronous or not ] On a journaled HFS+ volume, sync is synchronous because we flush the in-memory transaction buffer and the last thing that does is a flush-track-cache command which will not complete until all the other i/o's are complete. The behavior of other file systems is "mostly" synchronous. That is, they can issue async writes but after about 30 or so they will block on the queue of pending i/o going into the driver. In the worst case sync() could return with up to about 30 i/o's pending for a file system (on a local disk). Of course if you issue an fsync() on a file descriptor, that is *always* fully synchronous and will not return until the data is flushed. Having said all that, anyone that goes behind the back of the file system to read on disk data structures while the file system is still mounted is on thin ice. You can guarantee that we will break you in future releases. The functionality requested by the original poster is easily obtained with the PBCatSearch() interface. If you need sample code I can send it to you. It's a very fast way to search for files on an HFS file system. User level transactions would be very nice. It's not an easy problem though and definitely not something one can easily retrofit on existing file systems. As you surmise, writing a temp file, calling fsync() and then doing a rename() is the proper way to handle this. Of course user-level transactions are still a good thing but rename() (or the macos 9 style exchangedata()) give you most of what you need without the complexity of user level transactions. --dominic _______________________________________________ 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… Someday, I wish for a file system that is really a true database, with transactional semantics. They have transactional semantics in that each individual operation is atomic That's not even true. If you do this: create-write1-write2-write3-close and the system crashes, the result must be one of: .... I suppose you could write the file as a temp file, do a synchronous flush of the file, then move it into place. This is of course, a hack on top of an API that was never designed to be transactional. And you could still end up with a temp file lying around This email sent to site_archiver(a)lists.apple.com

1 0

Re: weird IOKit crash in IOIteratorNext()
by Ed Wynne 01 Dec '05

01 Dec '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Dec 1, 2005, at 4:57 PM, Julian Mayer wrote: 0 com.apple.CoreFoundation 0x901c0a14 CFRelease + 0x24 1 com.apple.framework.IOKit 0x939c0e24 IOIteratorNext + 0x14 more specifically the line that is crashing is: which seems very strange to me...i can see no obvious defect here... does anybody have any ideas when/why this crash could occur? _IOIteratorNext: +00 mfspr r0,lr +04 stw r0,0x8(r1) +08 stwu r1,0xffa0(r1) +0C addi r4,r1,0x40 +10 bl 0x905cd014 ; symbol stub for: _io_iterator_next +14 cmpwi cr7,r3,0x0 - its prebound, incorrectly. - its been stomped by a memory smasher. My bet, from having witnessed this happen before, is the first. -Ed _______________________________________________ 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… hello all i've received a very weird crash report that i'm unable to track down cause the code in question works everwhere else and similar code is also in apple's sample code. the crash is happening in: "for(object = IOIteratorNext(iter); object != 0; object = IOIteratorNext(iter)) // iterate through all "IOATABlockStorageDriver"s" If you look at the disassembled code for IOIteratorNext in IOKit.framework, you'll see there is no way this call sequence should happen. IOIteratorNext does not call CFRelease at + 0x10, it calls _io_iterator_next by way of a dyld_stub, which loads a function pointer and jumps to it. There are only two reasonable ways that dyld's function pointer could be wrong. This email sent to site_archiver(a)lists.apple.com

1 0

Re: weird IOKit crash in IOIteratorNext()
by j o a r 01 Dec '05

01 Dec '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On 1 dec 2005, at 22.57, Julian Mayer wrote: more specifically the line that is crashing is: which seems very strange to me...i can see no obvious defect here... does anybody have any ideas when/why this crash could occur? j o a r _______________________________________________ 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 "for(object = IOIteratorNext(iter); object != 0; object = IOIteratorNext(iter)) // iterate through all "IOATABlockStorageDriver"s" Perhaps you should also use "IOIteratorIsValid()"? The documentation states that: "Some iterators will be made invalid if changes are made to the structure they are iterating over. This function checks the iterator is still valid and should be called when IOIteratorNext returns zero." smime.p7s

1 0

Re: weird IOKit crash in IOIteratorNext()
by Julian Mayer 01 Dec '05

01 Dec '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com while( object = IOIteratorNext(iter) ) otherwise the for loop would never be entered at all thanks, julian _______________________________________________ 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… That's a pretty confusing looking for loop construct. You can get the same behavior out of this: It freaks me out that you have IOIteratorNext() in that first arguent to the for loop. I don't know anything about how and when each of the 3 expressions in the for loop is executed behind the scenes in compiler-land, but what you're doing scares me :) well i think the code you suggested is completely equivalent to mine but i will try to determine if the crash still occurs using your version... This email sent to site_archiver(a)lists.apple.com

1 0

Re: weird IOKit crash in IOIteratorNext()
by Ethan Bold 01 Dec '05

01 Dec '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com while( object = IOIteratorNext(iter) ) -ethan On Dec 1, 2005, at 1:57 PM, Julian Mayer wrote: 0 com.apple.CoreFoundation 0x901c0a14 CFRelease + 0x24 1 com.apple.framework.IOKit 0x939c0e24 IOIteratorNext + 0x14 more specifically the line that is crashing is: which seems very strange to me...i can see no obvious defect here... does anybody have any ideas when/why this crash could occur? please help ;) thanks, julian _______________________________________________ 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/ebold%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… That's a pretty confusing looking for loop construct. You can get the same behavior out of this: It freaks me out that you have IOIteratorNext() in that first arguent to the for loop. I don't know anything about how and when each of the 3 expressions in the for loop is executed behind the scenes in compiler-land, but what you're doing scares me :) hello all i've received a very weird crash report that i'm unable to track down cause the code in question works everwhere else and similar code is also in apple's sample code. the crash is happening in: "for(object = IOIteratorNext(iter); object != 0; object = IOIteratorNext(iter)) // iterate through all "IOATABlockStorageDriver"s" you can obtain the complete code in question here: http:// homepage.mac.com/julianmayer/SMARTReporterSource.zip and the complete crash log here: http://homepage.mac.com/ julianmayer/SMARTReporter.log and the system profile of the machine where the crash occurs here: http://homepage.mac.com/julianmayer/Profile.zip This email sent to ebold(a)apple.com This email sent to site_archiver(a)lists.apple.com

1 0

weird IOKit crash in IOIteratorNext()
by Julian Mayer 01 Dec '05

01 Dec '05

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com 0 com.apple.CoreFoundation 0x901c0a14 CFRelease + 0x24 1 com.apple.framework.IOKit 0x939c0e24 IOIteratorNext + 0x14 more specifically the line that is crashing is: which seems very strange to me...i can see no obvious defect here... does anybody have any ideas when/why this crash could occur? please help ;) thanks, julian _______________________________________________ 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… hello all i've received a very weird crash report that i'm unable to track down cause the code in question works everwhere else and similar code is also in apple's sample code. the crash is happening in: "for(object = IOIteratorNext(iter); object != 0; object = IOIteratorNext(iter)) // iterate through all "IOATABlockStorageDriver"s" you can obtain the complete code in question here: http:// homepage.mac.com/julianmayer/SMARTReporterSource.zip and the complete crash log here: http://homepage.mac.com/julianmayer/ SMARTReporter.log and the system profile of the machine where the crash occurs here: http://homepage.mac.com/julianmayer/Profile.zip This email sent to site_archiver(a)lists.apple.com

1 0