January 2012 - Darwin-dev - lists.apple.com

Re: recursive mutexes in Lion?
by Stephen J. Butler 31 Jan '12

31 Jan '12

On Tue, Jan 31, 2012 at 3:26 PM, Greg Parker <gparker(a)apple.com> wrote: > Why is the call to pthread_mutexattr_init() commented out? Does it work when you call pthread_mutexattr_init() first? > > I would not be surprised if pthread_mutexattr_settype() fails with EINVAL when given an uninitialized attr variable. Yeah, I saw the same issue with the sample code you did. But if you read the GCC bugzilla entry it appears there's a legitamate bug (or undocumented behavior) here. From what I gather in the comments, the problem isn't with the implementation of the pthread routines, but with the static initializers in certain conditions. _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.ap… This email sent to site_archiver(a)lists.apple.com

1 0

Re: recursive mutexes in Lion?
by Greg Parker 31 Jan '12

31 Jan '12

On Jan 31, 2012, at 1:26 PM, Greg Parker <gparker(a)apple.com> wrote: > On Jan 31, 2012, at 11:58 AM, Jack Howarth <howarth(a)bromo.med.uc.edu> wrote: >> The upcoming FSF gcc 4.7 release seems to have uncovered that the claimed >> recursive mutex support in pthread.h for Lion is broken... >> >> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51906 >> >> ---- pthread_test.C ---- >> #include <pthread.h> >> #include <stdio.h> >> >> /* int pthread_mutexatttr_settype(pthread_mutexattr_t *attr, int type); */ >> >> int main() >> { >> /* or PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP */ >> pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER; >> pthread_mutexattr_t mta; >> >> /* pthread_mutexattr_init(&mta); */ >> /* or PTHREAD_MUTEX_RECURSIVE_NP */ >> printf(" returned %d\n",pthread_mutexattr_settype(&mta, PTHREAD_MUTEX_RECURSIVE)); >> printf(" returned %d\n",pthread_mutex_trylock(&mutex)); >> } > > Why is the call to pthread_mutexattr_init() commented out? Does it work when you call pthread_mutexattr_init() first? > > I would not be surprised if pthread_mutexattr_settype() fails with EINVAL when given an uninitialized attr variable. Turns out there is a bug here that the gcc tests uncovered, but the code above doesn't show it. The problem is that pthread_mutex_trylock() does not recognize PTHREAD_RECURSIVE_MUTEX_INITIALIZER. If you use that static initializer and the first operation on the mutex is pthread_mutex_trylock() then it will fail with EINVAL. A recursive mutex using pthread_mutex_init() works. A recursive mutex for which the first operation is pthread_mutex_lock() also works. % clang test.c && ./a.out pthread_mutex_trylock mutex1 returned 22 % cat test.c #include <pthread.h> #include <stdio.h> int main() { int err; // fails: pthread_mutex_trylock of statically-initialized recursive mutex pthread_mutex_t mutex1 = PTHREAD_RECURSIVE_MUTEX_INITIALIZER; err = pthread_mutex_trylock(&mutex1); if (err) printf("pthread_mutex_trylock mutex1 returned %d\n", err); // succeeds: pthread_mutex_lock of statically-initialized recursive mutex pthread_mutex_t mutex2 = PTHREAD_RECURSIVE_MUTEX_INITIALIZER; err = pthread_mutex_lock(&mutex2); if (err) printf("pthread_mutex_lock mutex2 returned %d\n", err); } -- Greg Parker gparker(a)apple.com Runtime Wrangler _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.ap… This email sent to site_archiver(a)lists.apple.com

1 0

Re: recursive mutexes in Lion?
by Shantonu Sen 31 Jan '12

31 Jan '12

What are you hoping the test program to exercise, given that you're not initializing the mutexattr, nor are you associating the mutexattr with the mutex in any way? Using the compile-time mutex initializer gets you a default non-recursive mutex unless you use PTHREAD_RECURSIVE_MUTEX_INITIALIZER, which you're not using. This test program is really difficult to decipher. The following using either the proper static initializer or the proper mutexattr both seem to work for me. Are you seeing otherwise? $ cat test.c #include <pthread.h> #include <stdio.h> #include <err.h> #include <errno.h> int main(int argc, char *argv[]) { int result; #if USE_MUTEXATTR pthread_mutex_t mutex; pthread_mutexattr_t mta; result = pthread_mutexattr_init(&mta); if (result) err(1, "pthread_mutexattr_init"); result = pthread_mutexattr_settype(&mta, PTHREAD_MUTEX_RECURSIVE); if (result) err(1, "pthread_mutexattr_settype"); result = pthread_mutex_init(&mutex, &mta); if (result) err(1, "pthread_mutex_init"); result = pthread_mutexattr_destroy(&mta); if (result) err(1, "pthread_mutexattr_destroy"); #else pthread_mutex_t mutex = PTHREAD_RECURSIVE_MUTEX_INITIALIZER; #endif result = pthread_mutex_lock(&mutex); if (result) err(1, "pthread_mutex_lock 1"); result = pthread_mutex_lock(&mutex); if (result) err(1, "pthread_mutex_lock 2"); result = pthread_mutex_trylock(&mutex); if (result) { if (errno != EBUSY) { err(1, "pthread_mutex_trylock"); } } else { result = pthread_mutex_unlock(&mutex); if (result) err(1, "pthread_mutex_unlock trylock"); } result = pthread_mutex_unlock(&mutex); if (result) err(1, "pthread_mutex_unlock 2"); result = pthread_mutex_unlock(&mutex); if (result) err(1, "pthread_mutex_unlock 1"); result = pthread_mutex_destroy(&mutex); if (result) err(1, "pthread_mutex_destroy"); return 0; } $ cc -g -O0 test.c $ ./a.out $ cc -g -O0 test.c -DUSE_MUTEXATTR=1 $ ./a.out $ Shantonu On Jan 31, 2012, at 11:58 AM, Jack Howarth <howarth(a)bromo.med.uc.edu> wrote: > The upcoming FSF gcc 4.7 release seems to have uncovered that the claimed > recursive mutex support in pthread.h for Lion is broken... > > http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51906 > > WARNING: program timed out. > FAIL: 30_threads/lock/3.cc execution test > FAIL: 30_threads/recursive_mutex/try_lock/1.cc execution test > FAIL: 30_threads/try_lock/3.cc execution test > > Recursive mutexes are used due to the presence of... > > #if (!defined(_POSIX_C_SOURCE) && !defined(_XOPEN_SOURCE)) || defined(_DARWIN_C_SOURCE) > #define PTHREAD_ERRORCHECK_MUTEX_INITIALIZER {_PTHREAD_ERRORCHECK_MUTEX_SIG_init, {0}} > #define PTHREAD_RECURSIVE_MUTEX_INITIALIZER {_PTHREAD_RECURSIVE_MUTEX_SIG_init, {0}} > #endif /* (!_POSIX_C_SOURCE && !_XOPEN_SOURCE) || _DARWIN_C_SOURCE */ > > in /usr/include/pthread.h on Lion (but on SL which lacks that pthread.h change). > While a small test program like... > > ---- pthread_test.C ---- > #include <pthread.h> > #include <stdio.h> > > /* int pthread_mutexatttr_settype(pthread_mutexattr_t *attr, int type); */ > > int main() > { > /* or PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP */ > pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER; > pthread_mutexattr_t mta; > > /* pthread_mutexattr_init(&mta); */ > /* or PTHREAD_MUTEX_RECURSIVE_NP */ > printf(" returned %d\n",pthread_mutexattr_settype(&mta, PTHREAD_MUTEX_RECURSIVE)); > printf(" returned %d\n",pthread_mutex_trylock(&mutex)); > } > --------------------------- > > which can be compiled gcc-4.6 using "g++ -std=c++0x -g -O0 pthread_test.c -lpthread" on linux and produces... > > returned 0 > returned 0 > > it fails on darwin11 when using clang++ with " clang++ -std=c++0x -g -O0 pthread_test.c" where the resulting > binary produces... > > returned 22 > returned 0 > > suggesting that PTHREAD_MUTEX_RECURSIVE isn't a valid type. Is this breakage expected for lion? > Jack > _______________________________________________ > 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: > https://lists.apple.com/mailman/options/darwin-dev/ssen%40apple.com > > This email sent to ssen(a)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: https://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.ap… This email sent to site_archiver(a)lists.apple.com

1 0

Re: recursive mutexes in Lion?
by Greg Parker 31 Jan '12

31 Jan '12

On Jan 31, 2012, at 11:58 AM, Jack Howarth <howarth(a)bromo.med.uc.edu> wrote: > The upcoming FSF gcc 4.7 release seems to have uncovered that the claimed > recursive mutex support in pthread.h for Lion is broken... > > http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51906 > > ---- pthread_test.C ---- > #include <pthread.h> > #include <stdio.h> > > /* int pthread_mutexatttr_settype(pthread_mutexattr_t *attr, int type); */ > > int main() > { > /* or PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP */ > pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER; > pthread_mutexattr_t mta; > > /* pthread_mutexattr_init(&mta); */ > /* or PTHREAD_MUTEX_RECURSIVE_NP */ > printf(" returned %d\n",pthread_mutexattr_settype(&mta, PTHREAD_MUTEX_RECURSIVE)); > printf(" returned %d\n",pthread_mutex_trylock(&mutex)); > } Why is the call to pthread_mutexattr_init() commented out? Does it work when you call pthread_mutexattr_init() first? I would not be surprised if pthread_mutexattr_settype() fails with EINVAL when given an uninitialized attr variable. -- Greg Parker gparker(a)apple.com Runtime Wrangler _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.ap… This email sent to site_archiver(a)lists.apple.com

1 0

recursive mutexes in Lion?
by Jack Howarth 31 Jan '12

31 Jan '12

The upcoming FSF gcc 4.7 release seems to have uncovered that the claimed recursive mutex support in pthread.h for Lion is broken... http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51906 WARNING: program timed out. FAIL: 30_threads/lock/3.cc execution test FAIL: 30_threads/recursive_mutex/try_lock/1.cc execution test FAIL: 30_threads/try_lock/3.cc execution test Recursive mutexes are used due to the presence of... #if (!defined(_POSIX_C_SOURCE) && !defined(_XOPEN_SOURCE)) || defined(_DARWIN_C_SOURCE) #define PTHREAD_ERRORCHECK_MUTEX_INITIALIZER {_PTHREAD_ERRORCHECK_MUTEX_SIG_init, {0}} #define PTHREAD_RECURSIVE_MUTEX_INITIALIZER {_PTHREAD_RECURSIVE_MUTEX_SIG_init, {0}} #endif /* (!_POSIX_C_SOURCE && !_XOPEN_SOURCE) || _DARWIN_C_SOURCE */ in /usr/include/pthread.h on Lion (but on SL which lacks that pthread.h change). While a small test program like... ---- pthread_test.C ---- #include <pthread.h> #include <stdio.h> /* int pthread_mutexatttr_settype(pthread_mutexattr_t *attr, int type); */ int main() { /* or PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP */ pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER; pthread_mutexattr_t mta; /* pthread_mutexattr_init(&mta); */ /* or PTHREAD_MUTEX_RECURSIVE_NP */ printf(" returned %d\n",pthread_mutexattr_settype(&mta, PTHREAD_MUTEX_RECURSIVE)); printf(" returned %d\n",pthread_mutex_trylock(&mutex)); } --------------------------- which can be compiled gcc-4.6 using "g++ -std=c++0x -g -O0 pthread_test.c -lpthread" on linux and produces... returned 0 returned 0 it fails on darwin11 when using clang++ with " clang++ -std=c++0x -g -O0 pthread_test.c" where the resulting binary produces... returned 22 returned 0 suggesting that PTHREAD_MUTEX_RECURSIVE isn't a valid type. Is this breakage expected for lion? Jack _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.ap… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Reading HFS Standard Block0 on USB thumb drives
by Aaron Burghardt 30 Jan '12

30 Jan '12

Is this still an issue for you? Mac OS X supports mounting disks that have no partition map and the entire media is dedicated to a file system, although the system normally creates a partition map. In the case of HFS and HFS+, the volume header starts 1024 bytes into the partition (i.e., the third sector on a typical 512-bytes-per-sector hard disk), so try reading the first few blocks to confirm. Also, you can use Disk Utility to Get Info on the mounted partition to see which device is mounted. Disk Utility will show you the whole media object (which shows the partition map, if it exists) and the partitions on that media. Aaron On Jan 13, 2012, at 9:56 PM, Mike wrote: > Thanks Alison, > > I tried it on the device in question and it shows 512 bytes of zeros. > > Is your device HFS+ formatted or HFS? I am trying to read HFS only (as described in the old IM: Devices SCSI Manager chapter). > > I noticed 'diskUtil list' in Terminal shows that it doesn't recognize the partition format, yet the 10.6.8 Finder can mount it as read-only just fine and I see my files on it I copied over from when I formatted it with Drive Setup on OS 9: > > /dev/disk0 > #: TYPE NAME SIZE IDENTIFIER > 0: GUID_partition_scheme *1.0 TB disk0 > 1: EFI 209.7 MB disk0s1 > 2: Apple_HFS 10.6 569.4 GB disk0s2 > 3: Apple_HFS Files 269.3 GB disk0s3 > 4: Apple_HFS Applications 100.8 GB disk0s4 > 5: Apple_HFS 10.6 FCP X 51.3 GB disk0s5 > /dev/disk1 > #: TYPE NAME SIZE IDENTIFIER > 0: GUID_partition_scheme *80.0 GB disk1 > 1: EFI 209.7 MB disk1s1 > 2: Apple_HFS Files FireWire 79.7 GB disk1s2 > /dev/disk3 > #: TYPE NAME SIZE IDENTIFIER > 0: untitled *2.0 GB disk3 > > I know it's formatted correctly because I can see and access the files on it from Finder. > > Thanks, > > On 1/13/12 7:49 PM, Alison Cassidy wrote: >> I'm no expert, but have you tried running something like 'dd if=/dev/rdisk1 bs=512 count=1 | xxd' from the command-line? I just tried this on an old APM-formatted drive I have here and everything looks okay .... >> >> -- Allie >> >> On Jan 13, 2012, at 6:39 PM, Mike wrote: >> >>> > Hi, >>> > >>> > I am trying to read a Block0 driver descriptor map from old-style Apple Partition Map formatted USB thumb drives which were formatted using Mac OS Standard (HFS) on OS 9. I am using Xcode 4.2 on 10.6.8. >>> > >>> > But when I do the read the entire Block0 structure comes back filled with zeros. I've checked my device open and seek calls and they correctly set to byte offset 0 on the device. >>> > >>> > Any idea why this is happening? >>> > >>> > Thanks, > > _______________________________________________ > 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/aaron.burghardt%40gmail.c… > > This email sent to aaron.burghardt(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: https://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.ap… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Method to be notified of IP address changes
by Peter Speck 29 Jan '12

29 Jan '12

On 28/01/2012, at 23:33, Justin C. Walker wrote: > Also, see scutil(8). This provides a command-line interface to the functionality of the SystemConfiguration framework. The source is available on Apple's Darwin site, so you can see how you might use the APIs. Another example here: http://glimmerblocker.org/browser/trunk/PrefsPane/proxyApp/NetworkConfigura… This implements a listener using the API. This file implements a cmd line tool which prints the configuration to stdout whenever the configuration changes. ---- - Peter Speck _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.ap… This email sent to site_archiver(a)lists.apple.com

1 0

RE: Method to be notified of IP address changes
by Mark E. Armstrong 28 Jan '12

28 Jan '12

Thanks Justin... An example is just what the doctor ordered! > -----Original Message----- > From: darwin-dev-bounces @lists.apple.com > [mailto:darwin-dev-bounces @lists.apple.com] > On Behalf Of Justin > Sent: Saturday, January 28, 2012 2:33 PM > To: Darwin development mailing list > Subject: Re: Method to be notified of IP address changes > > > On Jan 28, 2012, at 12:45 , Jens Alfke wrote: > > > > > On Jan 27, 2012, at 5:50 PM, Mark E. Armstrong wrote: > > > >> I am looking for a way to be notified of an IP address change, > synchronously or asynchronously. What is the equivalent to > NotifyAddrChange in Windows or NETLINK in Linux? > > > > Use the SystemConfiguration framework to register for notifications of > changes to the status of a network interface. > > Also, see scutil(8). This provides a command-line interface to the > functionality of the SystemConfiguration framework. The source is > available on Apple's Darwin site, so you can see how you might use the > APIs. > > HTH > > Justin > > -- > Justin C. Walker, Curmudgeon at Large > Director > Institute for the Enhancement of the Director's income > ----------- > Question 43: > What if the hokey pokey > really *is* what it's all about? > -- > > > _______________________________________________ > 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: > https://lists.apple.com/mailman/options/darwin- > dev/mark.armstrong%40soleratec.com > > This email sent to mark.armstrong(a)soleratec.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: https://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.ap… This email sent to site_archiver(a)lists.apple.com

1 0

RE: Method to be notified of IP address changes
by Mark E. Armstrong 28 Jan '12

28 Jan '12

Thanks Jens... > -----Original Message----- > From: Jens > Sent: Saturday, January 28, 2012 12:45 PM > To: Mark E. Armstrong > Cc: darwin-dev(a)lists.apple.com > Subject: Re: Method to be notified of IP address changes > > > On Jan 27, 2012, at 5:50 PM, Mark E. Armstrong wrote: > > > I am looking for a way to be notified of an IP address change, > synchronously or asynchronously. What is the equivalent to > NotifyAddrChange in Windows or NETLINK in Linux? > > Use the SystemConfiguration framework to register for notifications of > changes to the status of a network interface. > > -Jens _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.ap… This email sent to site_archiver(a)lists.apple.com

1 0

Re: Method to be notified of IP address changes
by Justin C. Walker 28 Jan '12

28 Jan '12

On Jan 28, 2012, at 12:45 , Jens Alfke wrote: > > On Jan 27, 2012, at 5:50 PM, Mark E. Armstrong wrote: > >> I am looking for a way to be notified of an IP address change, synchronously or asynchronously. What is the equivalent to NotifyAddrChange in Windows or NETLINK in Linux? > > Use the SystemConfiguration framework to register for notifications of changes to the status of a network interface. Also, see scutil(8). This provides a command-line interface to the functionality of the SystemConfiguration framework. The source is available on Apple's Darwin site, so you can see how you might use the APIs. HTH Justin -- Justin C. Walker, Curmudgeon at Large Director Institute for the Enhancement of the Director's income ----------- Question 43: What if the hokey pokey really *is* what it’s all about? -- _______________________________________________ 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: https://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.ap… This email sent to site_archiver(a)lists.apple.com

1 0