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

Re: Crash in LSOpenFromRefSpec
by Finlay Dobbie 21 Jul '07

21 Jul '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:reply-to:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=X3tRt91yzDcvtELBQ/hN0WNePdQwS7LWP90aJDnKfLrQXp7E0szSqYyT4iRmn6xnX9A8vQXXrAnIY+ShlCmT+AF/45RVWiuWxMOE/izrtzWwPEFerGWvt6VlCMGnWosRPXsZ16Q+Q/eD9xdKJeqkyfGlHj2RWTLHzSJAJnDRxsA= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:reply-to:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=Y+O1AI4jYTMom6W6H+RbEXbBH8VkbwkA8TrMsGvW2u63KslcKJlU8CWiju4LELufsyIxGS6Npi34aPUOvXxH3NX0hex08BpnOIv7zU8oZY6t6g3cXxs7PLllkSPeAzVh3tXOIkW9CBL7hInaSv6yvfmtCtMgqk9vUkf4D4GT9pU= Launch Services is not part of Darwin. Try carbon-dev. -- Finlay Hi All, I am getting problem with the following function call LSOpenFromRefSpec.This works fine for PowerPc but when the same code is run with gdb under Intel based mac (processor is i386) I get Program received signal EXC_BAD_INSTRUCTION, Illegal instruction/operand. 0x9082c73b in _CFReadBytesFromFile (). I tried the settings of -mpreferred-stack-boundary in Other C and C++ flags of the Xcode projects but the result remained same.Am I doing something silly? Thanks in advance. Regards -A _______________________________________________ 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/finlay.dobbie%40gmail.com This email sent to finlay.dobbie(a)gmail.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… On 21/07/07, Arnab Ganguly <aganguly01(a)gmail.com> wrote: This email sent to site_archiver(a)lists.apple.com

1 0

Crash in LSOpenFromRefSpec
by Arnab Ganguly 21 Jul '07

21 Jul '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=N3yqRM3PxbTPt7vjJM1xVyliRKMfH90J6hHJc2p01x7s3U4ggTWt6FzgNdVFgKTV+PFF7axIpq7ZoD78BRDB3Iq427HqvaHHSyqeP8Dp5U10plkAbxcZN3080Lqrg/7nlSQEHFBleUx/N5VqKRDNWjkkPh5PoxcLiuAV6RT08p4= 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=ImBg7DjtH6LVrCDN6mtjAa5ZKLm2myZH2KOW2KYEnANT/OEpVuTkwDWBt//tmjI33d8MY1565xBLSRr/FmtzFpkU90Z3zX9rtNiedo6adFayBYOSlMGjwHZuR22aITob3Z9YUs/I4wFbm2+OYN9LPNc4D8cOyFRQ6/+Tu6lYGeY= _______________________________________________ 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: filename or full path from file descriptor or inode
by David Hoerl 21 Jul '07

21 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com 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… ... The canonically correct way to do this would be to remember the names of the files you have opened at the time you opened them. Alternately, you could use lsof, and it would report the files open (by name) by your process. Thanks for this information, but I simplified the question. In reality its a Cocoa app that uses a lot of open source libraries, and so I'm trying to get the list of open files, so as to better understand who is doing what. Of course, if they are temp files, the names won't help much! This email sent to site_archiver(a)lists.apple.com

1 0

Re: filename or full path from file descriptor or inode
by Terry Lambert 21 Jul '07

21 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com -- 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 Jul 20, 2007, at 11:07 AM, David Hoerl wrote: I am trying to find out why my app has so many file descriptors open at some point. Is there a straight forward way to get an ascii string with the filename and/or path given just a file descriptor? I can fstat the descriptor, so I can get a inode. The canonically correct way to do this would be to remember the names of the files you have opened at the time you opened them. Alternately, you could use lsof, and it would report the files open (by name) by your process. This email sent to site_archiver(a)lists.apple.com

1 0

filename or full path from file descriptor or inode
by David Hoerl 20 Jul '07

20 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com 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 am trying to find out why my app has so many file descriptors open at some point. Is there a straight forward way to get an ascii string with the filename and/or path given just a file descriptor? I can fstat the descriptor, so I can get a inode. This email sent to site_archiver(a)lists.apple.com

1 0

RE: Kernel Recursive mutexes 10.4.9 ++
by Erez Kaplan 20 Jul '07

20 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Thread-index: AcfKKsyCPewgwVLFROCuZvOft5ZX1QAmReRv Thread-topic: Kernel Recursive mutexes 10.4.9 ++ _______________________________________________ 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 Panic while trying to access the datastructure parameter in ioctl call.
by Terry Lambert 19 Jul '07

19 Jul '07

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com 4) The most likely cause is that your length is bogus; I notice that your buffer "data" (which should not be a stack buffer, it should be allocated!) is 2K in size, but you don't do anything to ensure that your count "bytes" is > 0 and less than 2K, so you could easily be overflowing the buffer. Another possibility is that, since 2K is a heck of a lot of data in kernel terms (1/8th of your entire available stack!!!), you could be trashing the kernel stack, resulting in the panic (this is less likely than simply having a bogus length and smashing your own stack, but it's a possibility). Then it's probably a stack smasher. 5) Finally, if you are using a worker thread to do the work instead of doing it in the context of the process making the call (you should know whether or not you are using one - in fact, you should have told us this when you asked the question...), then it's trying to copy in from kernel_process, which is the wrong process to copy in from. Basically, all copies in or out MUST occur in the context of the process requesting the copy. -- 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 Jul 19, 2007, at 12:18 PM, JanakiRam wrote: Is this a restriction that we can't send the data more than 2k from user space to kernel space ? . If we need to send more data from user space to general kext , pls suggest ways to do it. No, this is because you are using a stack buffer instead of an allocated buffer. Auto variables are allocated on the stack (this is part of the ISO/ANSI C standard). If you want a large data area, and want to avoid the risk, allocate the memory instead of declaring it to be on the stack (i.e. declare a pointer, allocate the memory, copy into the allocated memory, free the memory when you're done using it). Even after putting the > 0 && < 2k cindition , its still reproducing. I've a user space daemon process which gets created from the main program using pthreads. From User space daemon process i'm calling the ioctl to kernel extension. Do you mean this way of doing it is not the proper way ? If you start a worker thread in the kernel specifically associated with your device driver, then you will not be able to copy in from user space to the kernel. It doesn't sound like you are doing this. 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 -- Terry Terry, How do I access pthread_t from kernel? > pthread_t self; > > self = pthread_self(); Cheers, Erez On Jul 19, 2007, at 10:26 AM, Terry Lambert wrote: 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 //====================== //====================== 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… You don't. These are libc calls, not kernel calls. For the kernel, you will need to use a thread_t and current_thread(), and you will need to depend on the Mach symbol set to get at the function (more information available in KEXT programming introduction at <http://developer.apple.com >. On Jul 19, 2007, at 5:22 AM, Erez Kaplan wrote: 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 Panic while trying to access the datastructure parameter in ioctl call.
by JanakiRam 19 Jul '07

19 Jul '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:cc:in-reply-to:mime-version:content-type:references; b=X9OpQwBw7pU2Y9pI/gt+Jv4/hiXvamYuUm6DXMXLX140MRqpuSkiJvLxgIql8AvgBeTjvG0yBWJandCBEeU5n/KMxHzgOpHgO3LyskZ0D/f+lX+ABeDgMJmrLGdJRm2YYtYM2a3iIiI2j6wPt+RRltY0u9L/9wJerUD3v6AF7Og= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:references; b=qIa++XQ2lIYK3pTBcsYUjBr0Zmo0PCCoy9DE+299azUBFnQGOoD5s8bc3VS4SJ74/iiKAUYZGbU4A61PC3EyyEvFmHvonDgD33n3S1UiS/a74wyPYI8mWD+/K5giL8AeQpP3OB8uWX5bHwQZJrUMGTYZ5z/h1NK2hy3JqyHH29I= _______________________________________________ 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 Recursive mutexes 10.4.9 ++
by Alexei Kosut 19 Jul '07

19 Jul '07

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Alexei 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/akosut%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… On Jul 19, 2007, at 12:26 AM, Terry Lambert wrote: 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. Erez doesn't mention what type of kernel code he's writing, but it might be worth pointing out that IOKit has a recursive mutex implementation (IORecursiveLock) that is available. (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 akosut(a)apple.com This email sent to site_archiver(a)lists.apple.com

1 0