September 2004 - Darwin-dev

Re: bind
by Sam Hart 17 Sep '04

17 Sep '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On 17 Sep 2004, at 15:46, Matt Watson wrote: yes it does, would this be better than -DBIND_8_COMPAT=1 ? S a m ------------------------------------------------------- Drawing on my fine command of language, I said nothing. _______________________________________________ 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… Does adding -lresolv to your link line help? This email sent to site_archiver(a)lists.apple.com

1 0

Re: bind
by Giuliano Gavazzi 17 Sep '04

17 Sep '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com At 3:31 pm +0100 2004/09/17, Sam Hart wrote: [...] I am not building bind9, I am building an app that uses /usr/include/resolv.h when it comes to linking it has an undefined symbol _res_9_init in /usr/include/resolv.h res_init is defined as res_9_init adding -DBIND_8_COMPAT=1 gets rid of the problem, so I have a solution I guess I'm just slightly confused by the fact that I have to do this and that I haven't seen anyone else mention it Giuliano _______________________________________________ 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 don't know, should it be reported as a bug ? It happened to me in the past, but I do not remember for what. BTW, usually it just needs to be defined, not to take any particular value. This email sent to site_archiver(a)lists.apple.com

1 0

Re: bind
by Matt Watson 17 Sep '04

17 Sep '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Does adding -lresolv to your link line help? matt. On Sep 17, 2004, at 7:31 AM, Sam Hart wrote: when it comes to linking it has an undefined symbol _res_9_init _______________________________________________ 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 not building bind9, I am building an app that uses /usr/include/resolv.h This email sent to site_archiver(a)lists.apple.com

1 0

Re: bind
by Sam Hart 17 Sep '04

17 Sep '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On 17 Sep 2004, at 15:17, j o a r wrote: ./configure make sudo make install sorry, I should have been more clear when it comes to linking it has an undefined symbol _res_9_init in /usr/include/resolv.h res_init is defined as res_9_init adding -DBIND_8_COMPAT=1 gets rid of the problem, so I have a solution I guess I'm just slightly confused by the fact that I have to do this and that I haven't seen anyone else mention it I don't know, should it be reported as a bug ? S a m ------------------------------------------------------- Drawing on my fine command of language, I said nothing. _______________________________________________ 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 never had any problems compiling the standard named/bind distribution for Mac OS X. It has worked since Mac OS X Server 1.x (ie. pre-darwin) and up to the current Mac OS X 10.3.5. Note that I've never tried to do it on a plain Darwin install, only on Mac OS X. I simply: But perhaps you're talking about something else? I am not building bind9, I am building an app that uses /usr/include/resolv.h This email sent to site_archiver(a)lists.apple.com

1 0

Re: bind
by j o a r 17 Sep '04

17 Sep '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com ./configure make sudo make install But perhaps you're talking about something else? j o a r On 2004-09-17, at 15.30, Sam Hart wrote: while porting, I came across a problem with linking an undefined symbol res_9_init this turns out to be from using /usr/include/resolv.h I have had this confirmed on OpenDarwin 7.2.1, 10.3.4 and 10.3.5, so I think it is safe to say all the Darwin 7 releases. The fix was to add -DBIND_8_COMPAT=1 to my compile flags. why do we have the header files if there is no lib to support it ? bind9 appears to be in the CVS I am a little surprised that more people haven't tripped over this one, am I missing something ? _______________________________________________ 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 I've never had any problems compiling the standard named/bind distribution for Mac OS X. It has worked since Mac OS X Server 1.x (ie. pre-darwin) and up to the current Mac OS X 10.3.5. Note that I've never tried to do it on a plain Darwin install, only on Mac OS X. I simply: it appears res_init() is defined as res_9_init which comes from bind9, however it doesn't seem to be implemented, hence the error. smime.p7s

1 0

bind
by Sam Hart 17 Sep '04

17 Sep '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com while porting, I came across a problem with linking an undefined symbol res_9_init this turns out to be from using /usr/include/resolv.h I have had this confirmed on OpenDarwin 7.2.1, 10.3.4 and 10.3.5, so I think it is safe to say all the Darwin 7 releases. The fix was to add -DBIND_8_COMPAT=1 to my compile flags. why do we have the header files if there is no lib to support it ? bind9 appears to be in the CVS I am a little surprised that more people haven't tripped over this one, am I missing something ? S a m ------------------------------------------------------- Drawing on my fine command of language, I said nothing. _______________________________________________ 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… it appears res_init() is defined as res_9_init which comes from bind9, however it doesn't seem to be implemented, hence the error. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Using MIG RPC between 2 threads inside a process
by Mike Smith 17 Sep '04

17 Sep '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Sep 16, 2004, at 12:02 PM, darwin-dev-request(a)lists.apple.com wrote: 'ownership'. That is your 'client' thread allocates the 'record', initialises it and enqueues it. It agrees that it is not allowed to deal with the record until it gets returned to it on it's own queue, the only function that is allowed to operate on it is an 'abort' function if that is meaninngful. Another problem for me is that the thread which has things to enqueue is a time-constraint thread. I guess locking the pthread_mutex can cause priority inversion problems that I can not afford. This is why using this inter-thread *asynchonous* RPC mechanism, i was thinking that I could also avoid the priority inversion problem. What makes you think that the kernel has a lockless queue mechanism? = 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… It works best when you combine the queue with the concept of There are a number of relatively asynchronous handoff techniques that you could use, however fundamentally you have to accept that if your real-time thread depends on other, non-real-time components, you cannot guarantee real-time behaviour. If your queueing thread is the RT thread, then you should use pthread_mutex_trylock to obtain the queue lock, and if the lock op fails put the queued object onto a private deferred list and attempt to queue it again next time. This will prevent you from blocking behind the queue just because the dequeuer holds the lock on an empty or almost-empty queue. At this point, your only potential blocker is the source of queue objects. This email sent to site_archiver(a)lists.apple.com

1 0

Re: Using MIG RPC between 2 threads inside a process
by Paul Forgey 16 Sep '04

16 Sep '04

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com On Sep 16, 2004, at 4:02 PM, Godfrey van der Linden wrote: _______________________________________________ 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… If you only have one consumer, there is a way to do this without a mutex in a realtime thread. Check out Claudio Taglienti "Thread-Safe Circular Queue", C/C++ Users Journal 22(6), June 2004. Priority inversions are always a problem, I didn't realise you were talking about a "real-time" thread. All I can say is that this technique has a good probability of success 'cause the lock is only held for very, very short periods AND when we get priority handoff working out in user land it will definitely be presented through the pthread_mutex. (This is not a promise, I do not work in that part of Apple that would implement this API, I'm just making a guess based on the in kernel capabilities of mutex_lock()) If you need TRUE priority hand-off between a pair of thread then you will need kernel support and you will have to pay the cost of a kernel transition for every attempt to take a lock, this is probably not desirable either, though that is basically the mach_ipc solution. Godfrey This email sent to site_archiver(a)lists.apple.com

1 0

Re: AFS / CodaFS on Panther?
by Garance A Drosihn 16 Sep '04

16 Sep '04

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hello, all ... I'd like to experiment with filesystems of a more distributed nature i.e. more than just NFS-type mounts, more like systems that run daemons on clients that support callbacks, distributed locking mechanisms, etc. http://www.OpenAFS.org/ includes a version of openafs for MacOS 10. Of course, you need to have some afs (or openafs) servers to point your clients at. If you want to do this on darwinOS (instead of the full-blown MacOS 10), then you probably need to compile openafs from the source repositories. -- Garance Alistair Drosehn = gad(a)gilead.netel.rpi.edu Senior Systems Programmer or gad(a)freebsd.org Rensselaer Polytechnic Institute or drosih(a)rpi.edu _______________________________________________ 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… At 3:19 PM -0700 9/16/04, jmzorko(a)mac.com wrote: This email sent to site_archiver(a)lists.apple.com

1 0

Re: Using MIG RPC between 2 threads inside a process
by Godfrey van der Linden 16 Sep '04

16 Sep '04

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com Godfrey On Sep 16, , at 15:53, Stéphane LETZ wrote: On 16 sept. 04, at 23:11, Godfrey van der Linden wrote: Oops, sorry I forgot to signal the condition in enqueue should read OK Thanks. Thanks Stephane Letz _______________________________________________ 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… Priority inversions are always a problem, I didn't realise you were talking about a "real-time" thread. All I can say is that this technique has a good probability of success 'cause the lock is only held for very, very short periods AND when we get priority handoff working out in user land it will definitely be presented through the pthread_mutex. (This is not a promise, I do not work in that part of Apple that would implement this API, I'm just making a guess based on the in kernel capabilities of mutex_lock()) If you need TRUE priority hand-off between a pair of thread then you will need kernel support and you will have to pay the cost of a kernel transition for every attempt to take a lock, this is probably not desirable either, though that is basically the mach_ipc solution. void enqueue: (queue_entry_t) record { // Note no allocation to hold us up - it is assumed that the client has done the allocation itself pthread_mutex_lock(&queueLock); enqueue(&queueHead, record); pthread_cond_signal(&queueSleeper); pthread_mutex_unlock(&queueLock); } Also a small optimisation would be to track if you have any sleepers behind the dequeue and only do a signal if sleepers > 0. But what about the "priority inversion" problem? It a time-constraint writer thread is blocked on the "pthread_mutex_lock" done by a lower-priotity reader thread? This email sent to site_archiver(a)lists.apple.com

1 0