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

Re: PackageMaker command-line package size problem
by Stephane Sudre 19 Jul '07

19 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On 19 juil. 07, at 10:45, Paul Tapper wrote: Hi, _______________________________________________ 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've used PackageMaker, through its GUI, to make a pkg of a component, which came out as 765kB in size. To automate my build process a bit, I've now got a bash script which copys the component into a Temp folder and then builds the pkg using PackageMaker's command-line parameters, (extract below): "${PKGM}" -build -p ./MyAU.pkg -f ./Temp -d ./PackageStuff/AU/Description.plist -i ./PackageStuff/AU/info.plist This builds a pkg, which functions correctly. The only problem with it is that the pkg built with the command-line is 1.4mB in size. If I look at the contents, the difference in size is coming from the Contents/Resources sub-folder. In the GUI built pkg there's a .pax.gz file which is 1kB in size, but in the command-lne build pkg there's a .pax.gz file which is 700kB. Both pkg's have a .pax.gz of about 700kB in the Contents sub-folder, but it looks like the command-line built one has a duplicate in the Contents/Resources sub-folder. Does any one have any idea what's going on here? Depending on the package format, there can be a symbolic link in the Resources folder to the .pax.gz file in the Contents folder. If you're targeting Mac OS X 10.2 or later, you don't need this symbolic link. This email sent to site_archiver(a)lists.apple.com

1 0

Re: PackageMaker command-line package size problem
by Dirk Stegemann (Mailing-Lists) 19 Jul '07

19 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hello Paul, Am 19.07.2007 um 11:45 schrieb Paul Tapper: Hope ths helps, Dirk PS: Better to ask such questions on the Xcode-users mailing list. _______________________________________________ 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… To automate my build process a bit, I've now got a bash script which copys the component into a Temp folder and then builds the pkg using PackageMaker's command-line parameters, (extract below): "${PKGM}" -build -p ./MyAU.pkg -f ./Temp -d ./PackageStuff/AU/ Description.plist -i ./PackageStuff/AU/info.plist I'm building my installer package very similar, but I also provide a "-r /path/to/my/resources" argument. In the resulting pkg's "Resoures" folder, I see a "MyName.pax.gz" symlink file which points to the "Archive.pax.gz" file inside the "Contents" folder. Maybe you should specify a "-r" argument as well, to see how this changes PackageMaker's behaviour...? This email sent to site_archiver(a)lists.apple.com

1 0

Re: Kernel Recursive mutexes 10.4.9 ++
by Erez Kaplan 19 Jul '07

19 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.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… This email sent to site_archiver(a)lists.apple.com

1 0

PackageMaker command-line package size problem
by Paul Tapper 19 Jul '07

19 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hi, Does any one have any idea what's going on here? Thanks for any help Paul _______________________________________________ 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've used PackageMaker, through its GUI, to make a pkg of a component, which came out as 765kB in size. To automate my build process a bit, I've now got a bash script which copys the component into a Temp folder and then builds the pkg using PackageMaker's command-line parameters, (extract below): "${PKGM}" -build -p ./MyAU.pkg -f ./Temp -d ./PackageStuff/AU/Description.plist -i ./PackageStuff/AU/info.plist This builds a pkg, which functions correctly. The only problem with it is that the pkg built with the command-line is 1.4mB in size. If I look at the contents, the difference in size is coming from the Contents/Resources sub-folder. In the GUI built pkg there's a .pax.gz file which is 1kB in size, but in the command-lne build pkg there's a .pax.gz file which is 700kB. Both pkg's have a .pax.gz of about 700kB in the Contents sub-folder, but it looks like the command-line built one has a duplicate in the Contents/Resources sub-folder. _________________________________________________________________ Tell Hotmail about an email that changed your life! http://www.emailbritain.co.uk/ This email sent to site_archiver(a)lists.apple.com

1 0

Re: Kernel Recursive mutexes 10.4.9 ++
by Terry Lambert 19 Jul '07

19 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com I'd caution you on a couple of things: (1) your increments and decrements need to use atomic instructions -- Terry Terry, OK, got the idea of that solution. ======= Cheers, Erez On Jul 19, 2007, at 10:01 AM, Terry Lambert wrote: If funcB has to be callable unlocked somewhere, then: funcA() { lock mutex funcB_locked() do stuff... unlock mutex } funcB() { lock mutex funcB_locked() unlock mutex } /* MUST be called with mutex held! */ funcB_locked() { do stuff... } -- Terry Terry, Thanks for the explanation. funcA() { lock mutex funcB() do stuff... unlock mutex } funcB() { lock mutex do stuff... unlock mutex } Cheers, Erez On Jul 19, 2007, at 7:24 AM, Terry Lambert wrote: Hi, 10.4.9+ 10.5 How do I obtain a Kernel Recursive mutexe? The available api does not support this. lck_mtx_alloc_init(gmutex_grp_m, LCK_ATTR_NULL) some function(object) lock mutex atomic increment reference count unlock mutex ... long duration operation ... lock mutex atomic decrement reference count unlock mutex ... destroy(object) lock mutex again: /* Don't destroy the object while someone is using it */ if reference count msleep(mutex, object) goto again /* avoid race */ /* last reference gone, OK to kill it */ destroy object unlock mutex //====================== Erez Kaplan Senior software engineer OSX erezk(a)speedbit.com //====================== //====================== Erez Kaplan Senior software engineer OSX erezk(a)speedbit.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 avoided that solution because it fosters a reliance on recursive mutexes, and you tend to end up with subtle bugs in your code, if you depend on them. But yes, it's possible to build a recursive mutex construct from non-recursive mutexes. (2) the mutex should be held over an increment/decrement; if it's not, you will need to set up a retry, probably with a condition variable (3) it's possible to not hold the mutex over the decrement, but your objects would need to be create-once (it's probably too rough to go down that path, unless you are going to refactor your code anyway). (4) you are not going to be able to just declare static instances of these things; you will need an initializer function of some kind (minimally, you will need to initialize the interior mutex). On Jul 19, 2007, at 12:12 AM, Erez Kaplan wrote: I found another reference to this issue, is it a good one and does it cover my needs? > > typedef struct pthread_mutex_recursive_np_t { > pthread_mutex_t mutex; > pthread_t owner; > UInt32 count; > } pthread_mutex_recursive_np_t; > > void pthread_mutex_lock_recursive_np(pthread_mutex_recursive_np_t *mutex) > { > pthread_t self; > > self = pthread_self(); > if (mutex->owner != self) > { > pthread_mutex_lock(&mutex->mutex); > mutex->owner = self; > mutex->count = 1; > } > else > mutex->count += 1; > } > > void pthread_mutex_unlock_recursive_np(pthread_mutex_recursive_np_t > *mutex) > { > if (--mutex->count == 0) > { > mutex->owner = NULL; > pthread_mutex_unlock(&mutex->mutex); > } > } We do this sort of thing in the kernel in a number of places, where reference counts would be inappropriate. On Jul 18, 2007, at 9:36 PM, Erez Kaplan wrote: I am facing a some what strange sequence of calling (forced by PC porting) which looks something like this: Currently I get a deadlock situation. How would you outline handling this deadlock? On Jul 18, 2007, at 8:54 PM, Erez Kaplan wrote: On Jul 18, 2007, at 10:00 PM, Terry Lambert wrote: On Jul 18, 2007, at 2:37 AM, Erez Kaplan wrote: You don't; this is what reference and use counts are for. Can you explain a bit more or give reference to an example. For example, all thread reentrant FSs will take an I/O reference on a vnode, and use that reference to prevent the FS from being unmounted out from under an operation in progress, rather than holding a lock over the complete operation (thus unnecessarily serializing all other I/O to the vnode). Here is a simplified design pattern: /* wake up destroy routine; destroy routine may fail acquisition race */ if !reference count wakeup(object) -- Terry This email sent to site_archiver(a)lists.apple.com

1 0

Re: Kernel Recursive mutexes 10.4.9 ++
by Erez Kaplan 19 Jul '07

19 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.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… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Kernel Symbol Clarification
by Terry Lambert 19 Jul '07

19 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com int foo(int arg); you would: __private_extern__ int foo(int arg); -- Terry --- At Wed, 18 Jul 2007 22:59:55 -0700, Michael Smith wrote: Are you really sure about this? I thought the relocations were done by kextload in userspace, and the only namespaces that are shared are those of your kext and what it depends on. Drew, you are correct for more recent MacOS systems. Garth's description refers to a previous implementation. I would expect that all symbols that could be located within the kext would be satisfied first and then look for the rest of the symbols outside the kext. Am I not thinking about this correctly? ...Duane _______________________________________________ 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… Between objects in a KEXT, to prevent their visibility out5side to another KEXT that links against it. For example, say you have two KEXTs, and each, for whatever reason, links against the object file "bob.o"., and you also want KEXT1 to depend on KEXT2, so all the symbols in KEXT2 become visible to KEXT1. so in bob.o, to prevent, for example, the function from colliding when you tried to load KEXT1: both there and in the header. Then when the objects were relinked into a single relocatable object, the symbols would be removed from external visibility. This is also generally useful when you have an internal function and an external wrapper function, and you want to prevent people accidently calling the internal varsion of the function. For example, maybe the wrapper takes a mutex around the call, and internally, you know you will hold the mutex already over the call, but you don't want to make the mutex externally visible. On Jul 18, 2007, at 11:55 PM, Duane Murphy wrote: On Jul 18, 2007, at 12:03 PM, darwin-dev-request(a)lists.apple.com wrote: Garth Cummings writes: The kernel is one big monolithic namespace. All symbols in KEXTs must be unique across the system as they are linked into that one namespace. If a KEXT defines a symbol that already exists, the KEXT will not load. An alternative for symbols that are exported between objects in a kext, but which should not be visible outside it, is 'private extern'. Can you give an example of where private extern would help? I'm not seeing why this matters? If as Drew says, all linking is done except for external dependencies before being put into the kernel. This email sent to tlambert(a)apple.com This email sent to site_archiver(a)lists.apple.com

1 0

Re: Kernel Recursive mutexes 10.4.9 ++
by Terry Lambert 19 Jul '07

19 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com If funcB has to be callable unlocked somewhere, then: funcA() { lock mutex funcB_locked() do stuff... unlock mutex } funcB() { lock mutex funcB_locked() unlock mutex } /* MUST be called with mutex held! */ funcB_locked() { do stuff... } -- Terry Terry, Thanks for the explanation. funcA() { lock mutex funcB() do stuff... unlock mutex } funcB() { lock mutex do stuff... unlock mutex } Cheers, Erez On Jul 19, 2007, at 7:24 AM, Terry Lambert wrote: Hi, 10.4.9+ 10.5 How do I obtain a Kernel Recursive mutexe? The available api does not support this. lck_mtx_alloc_init(gmutex_grp_m, LCK_ATTR_NULL) some function(object) lock mutex atomic increment reference count unlock mutex ... long duration operation ... lock mutex atomic decrement reference count unlock mutex ... destroy(object) lock mutex again: /* Don't destroy the object while someone is using it */ if reference count msleep(mutex, object) goto again /* avoid race */ /* last reference gone, OK to kill it */ destroy object unlock mutex //====================== Erez Kaplan Senior software engineer OSX erezk(a)speedbit.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… We do this sort of thing in the kernel in a number of places, where reference counts would be inappropriate. On Jul 18, 2007, at 9:36 PM, Erez Kaplan wrote: I am facing a some what strange sequence of calling (forced by PC porting) which looks something like this: Currently I get a deadlock situation. How would you outline handling this deadlock? On Jul 18, 2007, at 8:54 PM, Erez Kaplan wrote: On Jul 18, 2007, at 10:00 PM, Terry Lambert wrote: On Jul 18, 2007, at 2:37 AM, Erez Kaplan wrote: You don't; this is what reference and use counts are for. Can you explain a bit more or give reference to an example. For example, all thread reentrant FSs will take an I/O reference on a vnode, and use that reference to prevent the FS from being unmounted out from under an operation in progress, rather than holding a lock over the complete operation (thus unnecessarily serializing all other I/O to the vnode). Here is a simplified design pattern: /* wake up destroy routine; destroy routine may fail acquisition race */ if !reference count wakeup(object) -- Terry This email sent to site_archiver(a)lists.apple.com

1 0

Re: Kernel Symbol Clarification
by Duane Murphy 19 Jul '07

19 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com --- At Wed, 18 Jul 2007 22:59:55 -0700, Michael Smith wrote: > >On Jul 18, 2007, at 12:03 PM, darwin-dev-request(a)lists.apple.com wrote: > >> Garth Cummings writes: >>> The kernel is one big monolithic namespace. All symbols in KEXTs must >>> be unique across the system as they are linked into that one >>> namespace. If a KEXT defines a symbol that already exists, the KEXT >>> will not load. >> >> Are you really sure about this? I thought the relocations were done >> by kextload in userspace, and the only namespaces that are shared >> are those of your kext and what it depends on. > >Drew, you are correct for more recent MacOS systems. > >Garth's description refers to a previous implementation. > >An alternative for symbols that are exported between objects in a >kext, but which should not be visible outside it, is 'private extern'. Can you give an example of where private extern would help? I'm not seeing why this matters? If as Drew says, all linking is done except for external dependencies before being put into the kernel. I would expect that all symbols that could be located within the kext would be satisfied first and then look for the rest of the symbols outside the kext. Am I not thinking about this correctly? ...Duane _______________________________________________ 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: Kernel Symbol Clarification
by Michael Smith 19 Jul '07

19 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Jul 18, 2007, at 12:03 PM, darwin-dev-request(a)lists.apple.com wrote: The kernel is one big monolithic namespace. All symbols in KEXTs must be unique across the system as they are linked into that one namespace. If a KEXT defines a symbol that already exists, the KEXT will not load. Are you really sure about this? I thought the relocations were done by kextload in userspace, and the only namespaces that are shared are those of your kext and what it depends on. Drew, you are correct for more recent MacOS systems. Garth's description refers to a previous implementation. = Mike _______________________________________________ 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… Garth Cummings writes: An alternative for symbols that are exported between objects in a kext, but which should not be visible outside it, is 'private extern'. This email sent to site_archiver(a)lists.apple.com

1 0