May 2008 - Darwin-dev - lists.apple.com

Re: Error code 0x0000c001 / 49153 during DADiskUnmount
by Jeff Johnson 23 May '08

23 May '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Tim, -Jeff On May 22, 2008, at 11:51 AM, Tim Schooley wrote: Hi folks, Apologies in advance if this is the wrong list (and don't hesitate to point me in the right direction). The code I'm writing needs to programmatically mount a device to a user-specified directory (such as /User/tester/mount_point). Since the FSMountLocalVolumeSync() and co. functions ignore the mount path (why, by the way?), I'm using the DA* functions. Both DADiskMountWithArguments() and DADiskUnmount() work correctly. Kind regards, Tim _______________________________________________ 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 believe that this is actually just an errno 1 EPERM "Operation not permitted". If you look at <DiskArbitration/DADissenter.h>, you have "typedef mach_error_t DAReturn". The format of a mach error is outlined in <mach/error.h>: it had 6 bits for the system, 12 bits for the subsystem, and 14 bits for the code. The system seems to be err_system (0x0), or err_kern. The subsystem seems to be err_sub(3). In the same header you have unix_error(errno) defined as "(err_kern|err_sub(3)| errno)". Then in <sys/errno.h> you have the error codes. The error itself might not be very helpful, but hopefully this explanation was. However, when I mount a device, then open a file in vim within the mount point, and then request an unmount, I recieve the following error code from the "dissenter" object: 0x0000c001. I am retrieving this error code using the normal DADissenterGetStatus() route. My problem is this: what is that code? It certainly isn't a DAReturn code, and I haven't been able to find any documentation on it. It is clear that it is a "device busy" kind of error (see http://www.shirt-pocket.com/forums/archive/index.php/t-689.html) but it'd be nice to know if there are any other "unknown" codes that're going to be returned in the unmount callback! The reason I am trying to unmount whilst having an open file within the mount point should be quite obvious - I need to handle cases where the mounted device is busy. Any help would be most appreciated. If you need extra info or code, just yell! :) This email sent to site_archiver(a)lists.apple.com

1 0

Re: DiskArbitration woes
by Ed Wynne 23 May '08

23 May '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.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: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com smime.p7s

1 0

Re: DiskArbitration woes
by Ron Aldrich 22 May '08

22 May '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.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: 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: DiskArbitration woes
by Ed Wynne 22 May '08

22 May '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Ron, -Ed On May 22, 2008, at 5:08 PM, Ron Aldrich wrote: Folks, Thanks for your time, Ron Aldrich Software Architects, Inc. _______________________________________________ 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/arwyn%40phasic.com This email sent to arwyn(a)phasic.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… This email sent to site_archiver(a)lists.apple.com This is usually a sign that the run loop upon which your DASession is registered upon is not running to service its event sources. I'm working on the disk formatting portion of a support application for a USB thumbdrive which supports hardware password protection, and I'm running into a couple of problems with Disk Arbitration. My problem occurs when I perform an operation that simply ejects the thumbdrive, changes it's state, and waits for the media to be remounted. Everything works as expected the first time, but the second time my diskPeeked callback is not called when the media remounts. I have one clue that might help explain it - There is an entry in the system log that reads: May 22 11:33:27 The-Dark-Side-2 diskarbitrationd[43]: SNG_Utility [337]:29515 not responding. Hunting around in the diskarbitrationd source, I found that this message is emitted by __DAResponseTimerCallback within DAQueue.c So, my question is: what sort of response is DiskArbitration looking for at this point? smime.p7s

1 0

DiskArbitration woes
by Ron Aldrich 22 May '08

22 May '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Folks, Thanks for your time, Ron Aldrich Software Architects, Inc. _______________________________________________ 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'm working on the disk formatting portion of a support application for a USB thumbdrive which supports hardware password protection, and I'm running into a couple of problems with Disk Arbitration. My problem occurs when I perform an operation that simply ejects the thumbdrive, changes it's state, and waits for the media to be remounted. Everything works as expected the first time, but the second time my diskPeeked callback is not called when the media remounts. I have one clue that might help explain it - There is an entry in the system log that reads: May 22 11:33:27 The-Dark-Side-2 diskarbitrationd[43]: SNG_Utility [337]:29515 not responding. Hunting around in the diskarbitrationd source, I found that this message is emitted by __DAResponseTimerCallback within DAQueue.c So, my question is: what sort of response is DiskArbitration looking for at this point? This email sent to site_archiver(a)lists.apple.com

1 0

Error code 0x0000c001 / 49153 during DADiskUnmount
by Tim Schooley 22 May '08

22 May '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com User-agent: Thunderbird 2.0.0.14 (X11/20080513) Hi folks, Apologies in advance if this is the wrong list (and don't hesitate to point me in the right direction). The code I'm writing needs to programmatically mount a device to a user-specified directory (such as /User/tester/mount_point). Since the FSMountLocalVolumeSync() and co. functions ignore the mount path (why, by the way?), I'm using the DA* functions. Both DADiskMountWithArguments() and DADiskUnmount() work correctly. However, when I mount a device, then open a file in vim within the mount point, and then request an unmount, I recieve the following error code from the "dissenter" object: 0x0000c001. I am retrieving this error code using the normal DADissenterGetStatus() route. The reason I am trying to unmount whilst having an open file within the mount point should be quite obvious - I need to handle cases where the mounted device is busy. Any help would be most appreciated. If you need extra info or code, just yell! :) Kind regards, Tim _______________________________________________ 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 My problem is this: what is that code? It certainly isn't a DAReturn code, and I haven't been able to find any documentation on it. It is clear that it is a "device busy" kind of error (see http://www.shirt-pocket.com/forums/archive/index.php/t-689.html) but it'd be nice to know if there are any other "unknown" codes that're going to be returned in the unmount callback! signature.asc signature.asc

1 0

Re: malloc_zone_batch_malloc
by Ariel Burton 22 May '08

22 May '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com User-agent: Mutt/1.4.1i Thanks. It seems odd. I tried another machine, and like your experience, it works. This is what I see on the two machines: i386-darwin-cereal % uname -a Darwin cereal .totalviewtech.com 8.4.1 Darwin Kernel Version 8.4.1: Tue Jan 3 18:23:53 PST 2006; root:xnu-792.7.90.obj~1/RELEASE_I386 i386 i386 386-darwin-cereal% ./t_mailing_list_repro do_experiment to size 16777216 returned 0 i386-darwin-wheat% uname -a Darwin wheat.totalviewtech.com 9.0.0 Darwin Kernel Version 9.0.0: Tue Oct 9 21:35:55 PDT 2007; root:xnu-1228~1/RELEASE_I386 i386 386-darwin-wheat% ./t_mailing_list_repro do_experiment to size 16777216 returned 1 What, I wonder, has changed? Ariel == On Wed 21 May 17:35, Eric Gouriou wrote: > > On May 21, 2008, at 5:03 PM, Ariel Burton wrote: > >The test program I included in my original > >psoting, however, tries all sizes from 1 bytes > >to 16mb without coming across a case where the > >routine returns a non-zero value. > > > >Am I missing something? > > Works for me on Leopard 10.5.2 . > > ------ > // cc -Wall --std=c99 -o mzm mzm.c > #include <stdio.h> > #include <malloc/malloc.h> > > static size_t > do_experiment (malloc_zone_t *z, size_t size_bound) > { > size_t ret_val = 0; > void *v[ 2 ]; > > for (size_t size = 1; size < size_bound; size++) > { > unsigned allocated = malloc_zone_batch_malloc ( z, size, v, 2 ); > if (allocated) { > ret_val = size; > break; > } > } > > return ret_val; > } /* do_experiment */ > > int > main (int argc, const char *argv []) > { > size_t size_bound = 16 * 1024 * 1024; > malloc_zone_t *z = malloc_default_zone(); > > size_t size = do_experiment(z, size_bound); > > printf("do_experiment to size %zd returned %zd\n", size_bound, size); > > return 0; > } > ------- > % cc -Wall --std=c99 -o mzm mzm.c > % ./mzm > do_experiment to size 16777216 returned 1 > ------- > > Of course I may have failed reading comprehension 201 this time > around... > > Eric > > >Ariel Burton > >== > >On Wed 21 May 16:37, Terry Lambert wrote: > >> > >>On May 21, 2008, at 2:00 PM, Ariel Burton wrote: > >>>In what circumstances is malloc_zone_batch_malloc > >>>supposed to return any memory? > >> > >><http://developer.apple.com/documentation/Performance/Conceptual/ManagingMem… > >>> > >> > >>"f your code allocates multiple, identically-sized memory blocks, you > >>can use the malloc_zone_batch_malloc function to allocate those > >>blocks > >>all at once. This function offers better performance than the > >>equivalent series of calls to malloc to allocate the same memory. > >>Performance is best when the individual block size is relatively > >>small— > >>less than 4K in size. The function does its best to allocate all of > >>the requested memory but may return less than was requested. When > >>using this function, check the return values carefully to see how > >>many > >>blocks were actually allocated." > >> > >> > >>-- 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/eric.gouriou%40pobox.com > > > >This email sent to eric.gouriou(a)pobox.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… This email sent to site_archiver(a)lists.apple.com

1 0

Re: malloc_zone_batch_malloc
by Eric Gouriou 22 May '08

22 May '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com The test program I included in my original psoting, however, tries all sizes from 1 bytes to 16mb without coming across a case where the routine returns a non-zero value. Works for me on Leopard 10.5.2 . ------ // cc -Wall --std=c99 -o mzm mzm.c #include <stdio.h> #include <malloc/malloc.h> static size_t do_experiment (malloc_zone_t *z, size_t size_bound) { size_t ret_val = 0; void *v[ 2 ]; for (size_t size = 1; size < size_bound; size++) { unsigned allocated = malloc_zone_batch_malloc ( z, size, v, 2 ); if (allocated) { ret_val = size; break; } } return ret_val; } /* do_experiment */ int main (int argc, const char *argv []) { size_t size_bound = 16 * 1024 * 1024; malloc_zone_t *z = malloc_default_zone(); size_t size = do_experiment(z, size_bound); printf("do_experiment to size %zd returned %zd\n", size_bound, size); return 0; } ------- % cc -Wall --std=c99 -o mzm mzm.c % ./mzm do_experiment to size 16777216 returned 1 ------- Eric Ariel Burton == On Wed 21 May 16:37, Terry Lambert wrote: In what circumstances is malloc_zone_batch_malloc supposed to return any memory? _______________________________________________ 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/eric.gouriou%40pobox.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 May 21, 2008, at 5:03 PM, Ariel Burton wrote: Am I missing something? Of course I may have failed reading comprehension 201 this time around... On May 21, 2008, at 2:00 PM, Ariel Burton wrote: <http://developer.apple.com/documentation/Performance/Conceptual/ManagingMem… "f your code allocates multiple, identically-sized memory blocks, you can use the malloc_zone_batch_malloc function to allocate those blocks all at once. This function offers better performance than the equivalent series of calls to malloc to allocate the same memory. Performance is best when the individual block size is relatively small— less than 4K in size. The function does its best to allocate all of the requested memory but may return less than was requested. When using this function, check the return values carefully to see how many blocks were actually allocated." -- Terry This email sent to eric.gouriou(a)pobox.com This email sent to site_archiver(a)lists.apple.com

1 0

Re: malloc_zone_batch_malloc
by Eric Gouriou 22 May '08

22 May '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com The test program I included in my original psoting, however, tries all sizes from 1 bytes to 16mb without coming across a case where the routine returns a non-zero value. Eric Eric Ariel Burton == On Wed 21 May 16:37, Terry Lambert wrote: In what circumstances is malloc_zone_batch_malloc supposed to return any memory? _______________________________________________ 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/eric.gouriou%40pobox.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 May 21, 2008, at 5:19 PM, Eric Gouriou wrote: On May 21, 2008, at 5:03 PM, Ariel Burton wrote: Am I missing something? Non-zero means success (or partial success) Oops, sorry I fail reading comprehension 101. Double-negatives are hard, let's go shopping. extern unsigned malloc_zone_batch_malloc(malloc_zone_t *zone, size_t size, void **results, unsigned num_requested); /* Allocates num blocks of the same size; Returns the number truly allocated (may be 0) */ On May 21, 2008, at 2:00 PM, Ariel Burton wrote: <http://developer.apple.com/documentation/Performance/Conceptual/ManagingMem… "f your code allocates multiple, identically-sized memory blocks, you can use the malloc_zone_batch_malloc function to allocate those blocks all at once. This function offers better performance than the equivalent series of calls to malloc to allocate the same memory. Performance is best when the individual block size is relatively small— less than 4K in size. The function does its best to allocate all of the requested memory but may return less than was requested. When using this function, check the return values carefully to see how many blocks were actually allocated." -- Terry This email sent to eric.gouriou(a)pobox.com This email sent to site_archiver(a)lists.apple.com

1 0

Re: malloc_zone_batch_malloc
by Eric Gouriou 22 May '08

22 May '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On May 21, 2008, at 5:03 PM, Ariel Burton wrote: The test program I included in my original psoting, however, tries all sizes from 1 bytes to 16mb without coming across a case where the routine returns a non-zero value. Non-zero means success (or partial success) Eric Ariel Burton == On Wed 21 May 16:37, Terry Lambert wrote: In what circumstances is malloc_zone_batch_malloc supposed to return any memory? _______________________________________________ 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/eric.gouriou%40pobox.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… Am I missing something? extern unsigned malloc_zone_batch_malloc(malloc_zone_t *zone, size_t size, void **results, unsigned num_requested); /* Allocates num blocks of the same size; Returns the number truly allocated (may be 0) */ On May 21, 2008, at 2:00 PM, Ariel Burton wrote: <http://developer.apple.com/documentation/Performance/Conceptual/ManagingMem… "f your code allocates multiple, identically-sized memory blocks, you can use the malloc_zone_batch_malloc function to allocate those blocks all at once. This function offers better performance than the equivalent series of calls to malloc to allocate the same memory. Performance is best when the individual block size is relatively small— less than 4K in size. The function does its best to allocate all of the requested memory but may return less than was requested. When using this function, check the return values carefully to see how many blocks were actually allocated." -- Terry This email sent to eric.gouriou(a)pobox.com This email sent to site_archiver(a)lists.apple.com

1 0