July 2004 - Darwin-kernel - lists.apple.com

Re: Regarding Kernel Code Coverage
by Justin Walker 08 Jul '04

08 Jul '04

On Jul 7, 2004, at 3:20, Manjunath wrote: Hi, I am trying to find the code coverage for kernel. Is it possible to find kernel code coverage for Darwin Kernel? I'm not sure what you mean by this. Usually, 'coverage' is determined by running a program against the source. Typically, doing this on kernel code is problematic, but I have seen related discussions on the FreeBSD lists. Can you be more specific about what you want? Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large * Institute for General Semantics | When LuteFisk is outlawed | Only outlaws will have | LuteFisk *--------------------------------------*-------------------------------* _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: Passing a char** to a Kext from user mode
by Jim Magee 08 Jul '04

08 Jul '04

See <bsm/audit.h> and related headers in the xnu/bsd/bsm directory for details on the audit subsystem in 10.3.4 (Darwin 7.4). What gets audited is any syscall with an authorization check in it (as well as several user-leve authorizations done by ssh, etc..). File accesses are pre-authorized through the open call, so the individual read() and write() calls are not audited. --Jim [demime 0.98b removed an attachment of type application/pkcs7-signature which had a name of smime.p7s] _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

FSMakeFSSpec, FSpMakeFSRef, FSRefMakePath
by Krishna Monian 07 Jul '04

07 Jul '04

Hi, I am trying to use the above function to generate the path for a given file. However I am facing the following problem. When I use call FSMakeFSSpec with the default vRef(0), the directory ID that I am interested in and 0 for the filename, the path is generated correctly. However, if I specify a filename which is present in the directory whose directory ID I pass in, it fails with -35 as the error code. Could anyone tell me what the reason behind this could be? Is there some special format for specifying the file name or is it just a null terminated string? Thanks Krishna __________________________________ Do you Yahoo!? New and Improved Yahoo! Mail - 100MB free storage! http://promotions.yahoo.com/new_mail _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: Mach Thread API
by Christoph Nägeli 07 Jul '04

07 Jul '04

On 7 Jul, 2004, at 17:23, Jim Magee wrote: Using Mach APIs directly to undermine the POSIX libraries is a good way to lead to memory leakage, corruption, and or crashes. In particular, there is no good way to use the Mach thread termination APIs without confusing the pthread layer. What is the thread doing when you want to kill it? If it is looping, can you add a call to pthread_testcancel()? If not, can you take a directed signal (pthread_kill()) and convert that into something you check for clean termination (after releasing locks, etc... held by the target thread). I'm porting a little othello game I had to write at a computer science course. So I got the main thread and one that is playing the game. If the user wants to play a new game then I have to kill the game thread (or at least restart it). => So the thread I want to kill is either listening on a condition 'pthread_cond_wait()' (Human, mouse click) or it is calculating some game-trees for AI. Because its variables are no more needed in the next game, I think it would be save to kill immediatly. >From your mail, I can see the following options. Correct me if I'm wrong. 1. Calling pthread_testcancel() during the thread. If I want to kill, call pthread_cancel followed by a broadcast on the condition variable in case the thread is waiting. I think it's a little annoying always calling pthread_testcancel() all the time. And If I do, how expensive would that be? 2. Forget about posix threads and work directly with Mach threads (thought they look a little complicated comparing to the pthreads => Is there any example code out there?). 3. Live with the memory leak (probably not the best solution) -- Christoph _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Volume Reference Numbers and directory IDs
by Krishna Monian 07 Jul '04

07 Jul '04

Hi, How can I find out what the default volume reference is and what they are for each of the mounted volumes? Another question I have is regarding the directory ID. Given a proc struct, how do I find the current directory ID from it? Currently I access the filedesc struct in the proc struc which in turn has a current current directory vnode. I use this to get the attributes using VOP_GETATTR and then access the va_fileid field in the vattr struct. Is this the right way to get the directory id? Thanks Krishna __________________________________ Do you Yahoo!? Yahoo! Mail - Helps protect you from nasty viruses. http://promotions.yahoo.com/new_mail _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: Mach Thread API
by Jim Magee 07 Jul '04

07 Jul '04

On Jul 6, 2004, at 7:59 AM, Christoph Ndgeli wrote: > Is there any documentation about the Mach Thread API? Yes, in the xnu source tree (download from Darwin) at xnu/osfmk/man/index.html > There are no man pages for the functions nor is there something in the > documentation > that came with the developer tools nor are the header files > documentated. The header files are MIG generated. But we normally ship the .defs IDL files which at least has some information (<mach/thread_act.defs>). > I need to kill a thread immediatly and as far as I can see, there is > no way to do it with posix threads > => I thought it would be pthread_cancel but that doesn't seem to work. The pthread_cancel() routine only interrupts a thread at a known clean point. POSIX defines many clean points, but only one is currently implemented in Mac OS X/Darwin (pthread_testcancel()). The reason is that the others are called throughout the upper layer frameworks of Mac OS X, but they do not implement cancellation safely logic yet...so it would be very unsafe to allow them in most Mac OS X applications. We are working on a way for pure POSIX applications to get the correct behavior while protecting Mac OS X-specific applications until the frameworks protect themselves. Stay tuned... Using Mach APIs directly to undermine the POSIX libraries is a good way to lead to memory leakage, corruption, and or crashes. In particular, there is no good way to use the Mach thread termination APIs without confusing the pthread layer. What is the thread doing when you want to kill it? If it is looping, can you add a call to pthread_testcancel()? If not, can you take a directed signal (pthread_kill()) and convert that into something you check for clean termination (after releasing locks, etc... held by the target thread). --Jim [demime 0.98b removed an attachment of type application/pkcs7-signature which had a name of smime.p7s] _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: Regarding Kernel Code Coverage
by Nikita Danilov 07 Jul '04

07 Jul '04

Justin Walker writes: > On Jul 7, 2004, at 23:30, Manjunath wrote: > > > > > On Thu, 2004-07-08 at 10:05, Justin Walker wrote: > >> On Jul 7, 2004, at 3:20, Manjunath wrote: > [snip] > > On linux we have Kernel Code Coverage Analysis where in we can measure > > how much of the kernel code is run during a test. GCC provides the code > > coverage at user level program if program is compiled with certain > > flags. On a similar basis can we get Code coverage for the kernel on > > Mac. > > I don't know that we have any tools that have been used for code > coverage. What tools do you use to get code coverage data for the > linux kernel? There is a patch that allows kernel to be compiled with -fprofile-arcs and -ftest-coverage. gcov input data (.bb files) are dynamically exported in /proc file system. > > Regards, > > Justin Nikita. _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: Mac throws "Can't fork" after forking daemon a number of times. W hats wrong ?
by Markus Hitter 07 Jul '04

07 Jul '04

Am 07.07.2004 um 12:38 schrieb Sanjay Mishra: I just pass all the arguments to it to one of the existing filters using execlp() and call wait() as the last function in the filter. You simulate the system() or popen() function ... After around 80 transactions and that too only on the single processor Mac machines it throws the error "can't fork". ... obviously less than perfect. What could be going wrong? Some table, noteably the kernels process table, might fill up. Try a "ps -auxww" while you notice the fault and look for something unusual. ... it gets the file from a predefined location ... (shuffling the data around through pipes should be faster and no disk space is needed) Are there any tool that can help me out. As a last resort, reduce code until it fits into an eMail and post it here. HTH, Markus - - - - - - - - - - - - - - - - - - - Dipl. Ing. Markus Hitter http://www.jump-ing.de/ _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Mac throws "Can't fork" after forking daemon a number of times. W hats wrong ?
by Sanjay Mishra 07 Jul '04

07 Jul '04

Hi, I have written a CUPS filter on Mac (Panther) and it is working fine on dual processor machines but on single processor Mac the OS throws an error "Can't fork" after around 80 transactions. Only when I am using my filter this message is thrown.I am not doing much in my filter.I just pass all the arguments to it to one of the existing filters using execlp() and call wait() as the last function in the filter. My application has many parts. One of them process the HTTP request. The other is the CUPS filter. The third which converts the document requested in the PDF format. Hence the sequence is that when the application receives the HTTP request for sending a file across it gets the file from a predefined location and converts it into PDF format (during this the CUPS filter gets called by the OS) and is sent across the network to the client. After around 80 transactions and that too only on the single processor Mac machines it throws the error "can't fork". What could be going wrong? Where I should start looking into ? I looked at the number of files my application opens and it is closing all of them at the proper place. Are there any tool that can help me out. Where ever I am forking I made sure that it does not hang and closes properly. I tested them.No error is thrown if I am running the application on dual processor system. Regards, Sanjay Mishra _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Regarding Kernel Code Coverage
by Manjunath 07 Jul '04

07 Jul '04

Hi, I am trying to find the code coverage for kernel. Is it possible to find kernel code coverage for Darwin Kernel? Regards Manju _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0