May 2006 - Darwin-dev - lists.apple.com

Re: POSIX shared memory question
by Brian Bergstrand 31 May '06

31 May '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On May 31, 2006, at 3:38 PM, Philip Lukidis wrote: ipcs thanks for any help, -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.3 (Darwin) iD8DBQFEfhLredHYW7bHtqIRAgF1AJsHAry8WGkZO2d8GpcFK9+xM0vS3ACgiS6n JiZHszJK6j9dPWIf1mnivmk= =eygN -----END PGP SIGNATURE----- _______________________________________________ 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… Hello. I have a question for using POSIX shared memory. There is a possibility that my process could crash before it would have called shm_unlink. I have two basic questions: a) What is the best method to deal with cleanup? I guess catching as many relevant signals as possible and cleaning up might help (SIGSEGV, SIGBUS, SIGABRT come to mind, though I'm not sure I could catch them), but is there a better approach, or complementary operations? There are at least two signals which I cannot catch. According to the manpage of shm_open, shared memory references do not survive a reboot, but I'd like to keep the filesystem as clean as possible. You could also create a destructor function and cleanup there, but either way it's still possible leave a dangling reference. You can't guarantee anything if you end up crashing. b) After I create the shared memory object via shm_open, call ftruncate, and map it via mmap, should I not see the object in the filesystem (well, after shm_open at least)? So far I cannot find it through the shell, yet no error is returned in my code. Is it created in a different namespace? Not unless you actually provide a valid path. Otherwise, an anonymous memory object is created in the kernel only with no representation in the FS. How can I view all created shared memory objects in the system? Unfortunately, that only works for SYSV objects. The POSIX ipcs equivalent is not included with OS X, so you are SOL there. Philip Lukidis This email sent to site_archiver(a)lists.apple.com

1 0

POSIX shared memory question
by Philip Lukidis 31 May '06

31 May '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Thread-index: AcaE8h0ZD1Nq/HCJSnScmCUh47uY/g== Thread-topic: POSIX shared memory question Hello. I have a question for using POSIX shared memory. There is a possibility that my process could crash before it would have called shm_unlink. I have two basic questions: a) What is the best method to deal with cleanup? I guess catching as many relevant signals as possible and cleaning up might help (SIGSEGV, SIGBUS, SIGABRT come to mind, though I'm not sure I could catch them), but is there a better approach, or complementary operations? There are at least two signals which I cannot catch. According to the manpage of shm_open, shared memory references do not survive a reboot, but I'd like to keep the filesystem as clean as possible. b) After I create the shared memory object via shm_open, call ftruncate, and map it via mmap, should I not see the object in the filesystem (well, after shm_open at least)? So far I cannot find it through the shell, yet no error is returned in my code. Is it created in a different namespace? How can I view all created shared memory objects in the system? thanks for any help, Philip Lukidis _______________________________________________ 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: forward-porting
by Michael Smith 31 May '06

31 May '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On May 31, 2006, at 8:32 AM, darwin-dev-request(a)lists.apple.com wrote: Send Darwin-dev mailing list submissions to darwin-dev(a)lists.apple.com Message: 6 Date: Wed, 31 May 2006 11:11:07 -0400 From: Morgan Winer <mhw23(a)cornell.edu> Subject: forward-porting To: darwin-dev(a)lists.apple.com Message-ID: <FE5B9978-827B-4CD9-8F7C-A392BCF2BBEC(a)cornell.edu> Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed I'm interested in porting 10.3.9 to work on the latest Apple hardware. Initial attempts show thats its likely something in a device driver, though which one is still beyond me. It hangs at different places, and at different times after power on. Any information (particular source files, packages even) would be of incredible help. This has already been done. = 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… The current version is "Mac OS X Version 10.4.6", and it ships with new Macintosh systems. This email sent to site_archiver(a)lists.apple.com

1 0

Re: DADiskEject and Resource Busy errors
by Dan Markarian 31 May '06

31 May '06

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: forward-porting
by Shawn Erickson 31 May '06

31 May '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=OY06yL6VTrpnFNyTUaCkLOuMDPqjJrjDJoTlTKAGS0vZO3pZjIW+Sx4ZPR1wZfn0B8R9n6JPl7S6iLfK+ccbcn1aZQy182ogDFewn6omKK2f3w9fT/DTvJ7IFkFfGIC1z8Q3fPhaTMOTffJCxMj4Vpk056VKXvJWW5Pexd1RdVA= I'm interested in porting 10.3.9 to work on the latest Apple hardware. As in Intel hardware? ...good luck :) I am curious on what the point and market for this would be? -Shawn _______________________________________________ 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 5/31/06, Morgan Winer <mhw23(a)cornell.edu> wrote: This email sent to site_archiver(a)lists.apple.com

1 0

Re: launchd, agents and tn2083
by Dave Zarzycki 31 May '06

31 May '06

site_archiver(a)lists.apple.com Delivered-To: Darwin-dev(a)lists.apple.com Stay tuned for Leopard though. :-) davez On May 30, 2006, at 5:20 PM, John Fieber wrote: tn2083 "Daemons and Agents" is a good read. Thanks, -john _______________________________________________ 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/zarzycki%40apple.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… You're mostly falling victim to a bug when it comes to the relationship between launchd and sessions. Please let it suffice for me to say that LaunchAgents don't work as advertised in Tiger. In the mean time, can you elaborate on what you're trying to do? Hopefully we can recommend an alternative solution in the mean time. Something isn't quite adding up though. Early on in the note: "Agents are useful because they can do things that daemons can't, like connect to the window server". But later on in Figure 5 (a decision tree for selecting an agent launch method) is a different story where by global login items (/Library/Preferences/ loginwindow.plist) are the only option for launching an agent for every user on login that can talk to the window server. But earlier in the document, that method is described as unsupported. launchd will work, but it also attempts to launch agents for ssh logins, which fails because the window server is inaccessible. (Of course, the agent could be fixed to detect that and deal with it sensibly, but in this particular organizational context I don't have control over the code in the agent.) So, there appears to be no *supported* way, according to tn2083 Figure 5, to launch an agent with GUI for all users on login. Am I reading that right? Or is there some undocumented launchd.plist option to suppress attempting to launch in a non-gui (sshd) context? This email sent to zarzycki(a)apple.com This email sent to site_archiver(a)lists.apple.com

1 0

Re: DKIOCCDREADTOC ioctl() fails during debugging
by Justin C. Walker 31 May '06

31 May '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On May 31, 2006, at 07:48 , Stephen F. Booth wrote: I have run into a rather strange problem I can't get my head around. Cheers, Justin -- Justin C. Walker, Curmudgeon-at-Large () The ASCII Ribbon Campaign /\ Help Cure HTML Email _______________________________________________ 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 have a program that works correctly when run from Xcode or Finder, but if I try to debug it a certain section of code fails every time. I am simply opening a CD using opendev() than calling ioctl() to read the disc's TOC: _fd = opendev((char *)[[self deviceName] fileSystemRepresentation], O_RDONLY | O_NONBLOCK, 0, NULL); if(-1 == ioctl([self fileDescriptor], DKIOCCDREADTOC, &cd_read_toc)) { I've omitted the error handling code, but I do checking. opendev() works every time, but as I mentioned, ioctl() fails from the debugger. Any ideas why? The system generally will tell you why something failed. I don't think anyone here is in a position to guess why. This email sent to site_archiver(a)lists.apple.com

1 0

Re: DKIOCCDREADTOC ioctl() fails during debugging
by Stephen F. Booth 31 May '06

31 May '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Stephen _______________________________________________ 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 31, 2006, at 8:36 AM, Garth Cummings wrote: I've omitted the error handling code, but I do checking. opendev () works every time, but as I mentioned, ioctl() fails from the debugger. Any ideas why? What's the value of errno when ioctl returns -1? That's the biggest clue about what "fails" means in this case. ;-) Ah, that would be helpful, wouldn't it! The error is 5:Input/output error. If you want to see the whole code in context it is on http:// svn.sourceforge.net/viewcvs.cgi/macaudiox/trunk/Drive/Drive.m? view=markup This email sent to site_archiver(a)lists.apple.com

1 0

Re: DKIOCCDREADTOC ioctl() fails during debugging
by Garth Cummings 31 May '06

31 May '06

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On May 31, 2006, at 7:48 AM, Stephen F. Booth wrote: --gc ____________________________________________________________________ Garth Cummings gcummings(a)apple.com Sr. Software Engineer Apple Developer Technical Support <http://developer.apple.com/technicalsupport> _______________________________________________ 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 omitted the error handling code, but I do checking. opendev() works every time, but as I mentioned, ioctl() fails from the debugger. Any ideas why? What's the value of errno when ioctl returns -1? That's the biggest clue about what "fails" means in this case. ;-) This email sent to site_archiver(a)lists.apple.com

1 0

Re: DADiskEject and Resource Busy errors
by Stephen F. Booth 31 May '06

31 May '06

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