May 2003 - Darwin-kernel - lists.apple.com

Re: Locking and synchronization primitives
by Wally Crooze 12 May '03

12 May '03

Quinn wrote: At 17:29 +1000 12/5/03, Wally Crooze wrote: Note. I am writing a filesystem... so I can't avoid using BSD headers. If you're working within the BSD portions of the kernel, your code should rely on the kernel funnels as its primary synchronisation mechanism. Kernel funnels are described in DTS Technote 2028 "Threading Architectures". <http://developer.apple.com/technotes/tn/tn2028.html#MacOSXKernelThreading> Any code running under a funnel is automatically single threaded. Thus, there's no need for locking for simple operations, like global += 1. However, locking does come into play if your code blocks. When you block you automatically release the funnel you're holding, and other code might run and modify state that you care about. You have to make sure that any state you depend on is preserved across blocking. Given that you're working on a file system, the primary locking mechanism you should use is the vnode locking mechanism (vn_lock--which calls VOP_LOCK--and VOP_UNLOCK). If you need to lock something other than vnodes, you would typically use the BSD locks provided by <sys/lock.h>. Remember that vnode locks must be taken in the correct order (from root to leaf) in order to avoid deadlocks. This system will be multithreaded as it is required to do work in the background and keep synchronized with each other as well as normal vfs / vnode requests. So I will have to look more into sys/lock.h and read the TechNote. Thanks -- Wally Crooze +-------------------------------------------------------+ | ,-_|\ Wally Crooze <wallycrooze(a)pobox.com> | | / \ | | \_,-._/ | | v | +-------------------------------------------------------+ _______________________________________________ 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: Locking and synchronization primitives
by Quinn 12 May '03

12 May '03

At 17:29 +1000 12/5/03, Wally Crooze wrote: Note. I am writing a filesystem... so I can't avoid using BSD headers. If you're working within the BSD portions of the kernel, your code should rely on the kernel funnels as its primary synchronisation mechanism. Kernel funnels are described in DTS Technote 2028 "Threading Architectures". <http://developer.apple.com/technotes/tn/tn2028.html#MacOSXKernelThreading> Any code running under a funnel is automatically single threaded. Thus, there's no need for locking for simple operations, like global += 1. However, locking does come into play if your code blocks. When you block you automatically release the funnel you're holding, and other code might run and modify state that you care about. You have to make sure that any state you depend on is preserved across blocking. Given that you're working on a file system, the primary locking mechanism you should use is the vnode locking mechanism (vn_lock--which calls VOP_LOCK--and VOP_UNLOCK). If you need to lock something other than vnodes, you would typically use the BSD locks provided by <sys/lock.h>. Remember that vnode locks must be taken in the correct order (from root to leaf) in order to avoid deadlocks. S+E -- Quinn "The Eskimo!" <http://www.apple.com/developer/> Apple Developer Technical Support * Networking, Communications, Hardware _______________________________________________ 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

Locking and synchronization primitives
by Wally Crooze 12 May '03

12 May '03

G'day, I've been investigating the different locking/synchronization primitives with respect to different OS kernels. In darwin it appears that there are generally two places to look for primitives... mach (ofsmk) or bsd. Reading the Kernel Programming guide the section Synchronization Primitives outlines the recommended primitives, which are defined by the mach. So... I attempted to use spinlocks with the simple_lock interface (defined in kern/simple_lock.h), but kept getting the complaint "passed 2 arguments, but takes just 1" for simple_lock_init. What I found was that a redefinition of the simple_lock interface was made in sys/lock.h and simple_lock and simple_unlock functions do nothing, which overrides kern/simple_lock.h. I can't avoid including this because it is brought in by sys/vnode.h, and many other headers. Does anybody know why simple_lock interfaced is redefined and why the lock and unlock do nothing? I need this to work on a multiprocessor machine... and this doesn't appear to help. Also, are the other interfaces for locking/synchronization described in Kernel Programming correct? Should I be using different interfaces in BSD from mach? What does monotreme mean? Note. I am writing a filesystem... so I can't avoid using BSD headers. Thanks. -- Wally Crooze +-------------------------------------------------------+ | ,-_|\ Wally Crooze <wallycrooze(a)pobox.com> | | / \ | | \_,-._/ | | v | +-------------------------------------------------------+ _______________________________________________ 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

Kernel Trace
by Joey Echeverria 10 May '03

10 May '03

Is there a method of tracing function calls in the kernel? I looked at ktrace, but that only gave me the syscalls and returns from syscalls. What I want to see is the "complete" chain of kernel calls. For example, on a read() I'd like to be able to see its interactions with filesystem layers, unified buffer cache and even the I/O Kit ATA drivers if possible. Any help would be greatly appreciated. -Joey Joseph G. Echeverria Department of Electrical and Computer Engineering Carnegie Mellon University _______________________________________________ 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: Threads
by Quinn 07 May '03

07 May '03

At 14:09 +0200 7/5/03, Thierry Passeron wrote: i just wondered if in OSX threads were user-threads or kernel-threads, i'm really new to threads and i came to think that since the API lets one change the schedule of a thread it has to be user-threads. can someone please tell me the answer ? There are many different types of threads on Mac OS X, so your question has no general answer. However, if you're talking about pthreads, then yes, a pthread is kernel thread. A pthread is actually a wrapper around a Mach thread, which is a kernel thread. This is explained in detail in DTS Technote 2028 "Threading Architectures". <http://developer.apple.com/technotes/tn/tn2028.html> S+E -- Quinn "The Eskimo!" <http://www.apple.com/developer/> Apple Developer Technical Support * Networking, Communications, Hardware _______________________________________________ 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

Threads
by Thierry Passeron 07 May '03

07 May '03

Hi everybody, i just wondered if in OSX threads were user-threads or kernel-threads, i'm really new to threads and i came to think that since the API lets one change the schedule of a thread it has to be user-threads. can someone please tell me the answer ? Best regards, Thierry. ps: i'm making yet an other dining philosophers program ... :) _______________________________________________ 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: aio?
by Steve Dekorte 06 May '03

06 May '03

On Tuesday, May 6, 2003, at 02:01 AM, Steve Dekorte wrote: Are there any plans to support asynchronous disk operations? See: http://www.ixora.com.au/notes/asynchronous_io.htm On 2nd thought, let me rephrase that: Is aio the sort of thing that the darwin kernel folks find interesting? (since Apple seems to have a policy of never discussing "plans" :-)) Cheers, Steve Io, a small language: http://www.iolanguage.com/ _______________________________________________ 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

aio?
by Steve Dekorte 06 May '03

06 May '03

Hi, Are there any plans to support asynchronous disk operations? See: http://www.ixora.com.au/notes/asynchronous_io.htm Cheers, Steve OSX freeware and shareware: http://www.dekorte.com/downloads.html _______________________________________________ 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: Building Darwin 6.5
by Paul Ripke 04 May '03

04 May '03

On Monday, May 5, 2003, at 07:52 Australia/Sydney, Thierry DELHAISE wrote: Like certainly many others, I'm trying to build Darwin 6.5 on my 10.2.5 with December developper tools 2002. <snip> [ updating libsa.o RELEASE ] /usr/bin/ld -static -r -arch ppc -o libsa.o bootstrap.cpo catalogue.cpo c++rem3.o kld_patch.o kmod.cpo mach.o misc.o mkext.o vers_rsrc.o bsearch.o malloc.o sort.o strrchr.o strstr.o setjmp.o vers.o /usr/local/lib/libkld.a /usr/bin/ld: table of contents for archive: /usr/local/lib/libkld.a is out of date; rerun ranlib(1) (can't load from it) make[6]: *** [libsa.o] Error 1 make[5]: *** [build_all] Error 2 make[4]: *** [do_all] Error 2 make[3]: *** [build_all] Error 2 make[2]: *** [build_all] Error 2 make[1]: *** [build_all] Error 2 make: *** [all] Error 2 Since 'libkld.a' is built from cctools/ld, I've rerun 'ranlib' on it (and make a new fresh copy in /usr/local/lib' But nothing works. I've allways the same message. I don't have much to add, except on my system where I've built many a kernel, including 10.2.5 (aka 6.5): ksh$ ls -l /usr/local/lib/libkld.a -rw-r--r-- 1 root wheel 862092 Jun 1 2001 /usr/local/lib/libkld.a ksh$ ar -tv /usr/local/lib/libkld.a ar: /usr/local/lib/libkld.a: Inappropriate file type or format ksh$ file /usr/local/lib/libkld.a /usr/local/lib/libkld.a: Mach-O object ppc As above, it's been a while since I'd built and installed it. It still works, but it doesn't appear to be a library. Cheers, -- Paul Ripke Unix/OpenVMS/TSM/DBA 101 reasons why you can't find your Sysadmin: 68: It's 9AM. He/She is not working that late. -- Koos van den Hout _______________________________________________ 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: Building Darwin 6.5
by Thierry DELHAISE 04 May '03

04 May '03

Le lundi, 5 mai 2003, ` 00:37 Europe/Paris, Michael A. Williams a icrit : "sudo ranlib /usr/local/lib/libkld.a" worked for me. Yes, you're right. Don't know why, but the first time I've tried it didn't solve the problem. May be to many "coffe" (it's a recommendation in "the Building Darwin" doc). May be I could submit a bug report about that ;-) THX On Sun, 4 May 2003, Thierry DELHAISE wrote: /usr/bin/ld: table of contents for archive: /usr/local/lib/libkld.a is out of date; rerun ranlib(1) (can't load from it) --Mike. _______________________________________________ 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. _______________________________________________ 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