August 2005 - Darwin-kernel

fsync()
by Steve Dekorte 15 Aug '05

15 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com -- Steve _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… I've been told that the docs for fync() are incorrect and that OSX does not actually wait for the data to be written to the disk before returning. Is this true and if so, how can I ensure the data is written before continuing? This email sent to site_archiver(a)lists.apple.com

1 0

Re: mbufs and cluster
by Michael Tuexen 15 Aug '05

15 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hi Josh, Best reagrds Michael On Aug 15, 2005, at 5:24 PM, Josh Graessley wrote: _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… thank you very much for the information, so m_expand from bsd/kern/uipc_mbuf.c is always used to allocate mbufs. If I remember correctly, mbufs themselves are stored in clusters. A cluster can be used as a cluster on it's own or it can be used to store mbufs. An mbuf is a fraction of a cluster. If the system runs out of mbufs, it may convert a cluster in to something like 8 mbufs. This email sent to site_archiver(a)lists.apple.com

1 0

Re: mbufs and cluster
by Josh Graessley 15 Aug '05

15 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com -josh On Aug 15, 2005, at 7:46 AM, Michael Tuexen wrote: Dear all, I have a simple question: But netstat -m gives for example: [PowerMac:/] tuexen% netstat -m 330 mbufs in use: 325 mbufs allocated to data 1 mbufs allocated to packet headers 3 mbufs allocated to socket names and addresses 1 mbufs allocated to Appletalk data blocks 376/848 mbuf clusters in use 1778 Kbytes allocated to network (46% in use) 0 requests for memory denied 0 requests for memory delayed 0 calls to protocol drain routines Thank you very much for your help. Best regards Michael This email sent to jgraessley(a)apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com If I remember correctly, mbufs themselves are stored in clusters. A cluster can be used as a cluster on it's own or it can be used to store mbufs. An mbuf is a fraction of a cluster. If the system runs out of mbufs, it may convert a cluster in to something like 8 mbufs. Normally a used cluster element is one where there exist an mbuf which has a pointer to the cluster element. This happens at least when you get it with m_mclget or the corresponding macro. This would than mean that the number of uses mbufs is not larger than the number of used mbuf clusters. Why can it happen that the number of mbuf clusters is larger than the number of mbufs? _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/jgraessley% 40apple.com smime.p7s

1 0

Re: mbufs and cluster
by Michael Cashwell 15 Aug '05

15 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com On Aug 15, 2005, at 10:46 AM, Michael Tuexen wrote: Dear all, I have a simple question: But netstat -m gives for example: [PowerMac:/] tuexen% netstat -m 330 mbufs in use: 325 mbufs allocated to data 1 mbufs allocated to packet headers 3 mbufs allocated to socket names and addresses 1 mbufs allocated to Appletalk data blocks 376/848 mbuf clusters in use 1778 Kbytes allocated to network (46% in use) 0 requests for memory denied 0 requests for memory delayed 0 calls to protocol drain routines Thank you very much for your help. Best regards Michael -Mike _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… Normally a used cluster element is one where there exist an mbuf which has a pointer to the cluster element. This happens at least when you get it with m_mclget or the corresponding macro. This would than mean that the number of uses mbufs is not larger than the number of used mbuf clusters. Why can it happen that the number of mbuf clusters is larger than the number of mbufs? If you see this GROW over time then it may just be buggy code that's leaking clusters. if it's pretty static then I have a theory: I could see this if a network driver allocated clusters for DMA-ready hardware that can't work on mbufs directly. I ran into this with a card that had the notion of 3 buffer sizes but always started a received frame (even large ones) with one of the small buffers followed by a chain of large ones. Given this strange chaining I couldn't given it mbufs directly (via mbuf memory descriptors) because the hardware would chain large buffers after a small one for large frames. This put the hardware chaining at the wrong level to use mbufs directly. Handing it bare clusters of various sizes and attaching them to an mbuf after the frames are received was the best fit I could come up with. It worked well but did produce the effect you are seeing here of mbuf-less clusters being allocated. If I'm right you are seeing memory waiting for DMA. This email sent to site_archiver(a)lists.apple.com

1 0

mbufs and cluster
by Michael Tuexen 15 Aug '05

15 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Dear all, I have a simple question: But netstat -m gives for example: [PowerMac:/] tuexen% netstat -m 330 mbufs in use: 325 mbufs allocated to data 1 mbufs allocated to packet headers 3 mbufs allocated to socket names and addresses 1 mbufs allocated to Appletalk data blocks 376/848 mbuf clusters in use 1778 Kbytes allocated to network (46% in use) 0 requests for memory denied 0 requests for memory delayed 0 calls to protocol drain routines Thank you very much for your help. Best regards Michael _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… Normally a used cluster element is one where there exist an mbuf which has a pointer to the cluster element. This happens at least when you get it with m_mclget or the corresponding macro. This would than mean that the number of uses mbufs is not larger than the number of used mbuf clusters. Why can it happen that the number of mbuf clusters is larger than the number of mbufs? This email sent to site_archiver(a)lists.apple.com

1 0

Re: Carbon File Manger VFS caching
by Quinn 15 Aug '05

15 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com <http://developer.apple.com/technotes/tn2004/tn2124.html> The above applies only to Posix volumes; IIRC that's what you're implementing. <http://developer.apple.com/samplecode/FileNotification/FileNotification.html> S+E -- Quinn "The Eskimo!" <http://www.apple.com/developer/> Apple Developer Technical Support * Networking, Communications, Hardware _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… At 18:37 -0500 12/8/05, Brian Bergstrand wrote: We've had problems with the Carbon File Mgr caching dir entries for too long in Panther and Jaguar and the problem persists in Tiger. Both the Finder and the File Manager cache directory contents. You need to distinguish between the two. The way to do that is to write a simple tool that iterates a directory using File Manager calls. If you're seeing stale data, the File Manager cache is the problem. OTOH, if the data is up-to-date, you need to look to the Finder. I believe that the enumeration cache is per-process. If so, your tool will to have to keep running between successive enumerations. You can dump the enumeration cache for a volume using PrintEnumerationCache, as described in Technote 2124 "Debugging Magic". As far as the File Manager is concerned, looking at the source, it seems that it caches directory contents based the modification time of the directory. To improve performance, it only rechecks the directory's modification time when a client starts an enumeration. You should make sure you're bumping the mod date of the directory when you make changes to the directory. If the Finder is the problem, I believe that, in 10.4 and later, the Finder will use kqueues to pick up changes to its windows. To see an example of (roughly) what the Finder does, take a look at the FileNotification sample. In theory, if your file system notifies the system of changes, this will automatically bubble up through kqueues to the Finder. AFAIK you can do this will public KPIs. For an example of how to do this, search the HFS [Plus] source for "HFS_KNOTE". In practice, this isn't going to work because, looking at the Finder source, it seems that it contains a special case list of file systems that support kqueues. *grrrr* <rdar://problem/4216129> This email sent to site_archiver(a)lists.apple.com

1 0

Re: Implementing Mutex in IOKIT driver
by Brian Bergstrand 13 Aug '05

13 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Aug 13, 2005, at 6:45 AM, santosh katvate wrote: Hello everybody I want to know how to implement mutex in case of driver. i am able to trap read routine in my driver(which has ioblockstoragedriver as its base class) but for some of the read request i want to do some processing for which i want that only after one request is completed ,that the second request should be processed Till that time other request should be in wait state. Thanking in advance Santosh k Brian Bergstrand <http://www.bergstrand.org/brian/> PGP Key ID: 0xB6C7B6A2 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) iD8DBQFC/iN8edHYW7bHtqIRAhP9AKD02Mpzp2z42Ntd9IvWFCOwTTTsZACeMYyZ NVvFBNgTSRndi6A543vV9GM= =3KzW -----END PGP SIGNATURE----- _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… You didn't specify what kernel version, so I'll assume 8. All of the primitive mach lock functions are declared in <kern/locks/h>. IOKit locks are just wrappers around these. What are the ways in which i can handle this situation. I am using pair of IOLOCK and IOUNLOCK function but it results in system hang This email sent to site_archiver(a)lists.apple.com

1 0

Implementing Mutex in IOKIT driver
by santosh katvate 13 Aug '05

13 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(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:mime-version:content-type:content-transfer-encoding:content-disposition; b=sgsDimv3ASZNr2ZTbctr1sN6EkIbOt/jPLT++UjNEZTDr39LNkICm/Q3HVijt2zenQbGNR1tCZ2GLslDBGbQtKahL94KKQbKpaL5l1VyajxA3szD1DevMlbTUpVRA3NubscDLKP0ohNFUtpyfCiP4sSKBNi+Up2yvrPy1D+DeTI= Hello everybody I want to know how to implement mutex in case of driver. i am able to trap read routine in my driver(which has ioblockstoragedriver as its base class) but for some of the read request i want to do some processing for which i want that only after one request is completed ,that the second request should be processed Till that time other request should be in wait state. What are the ways in which i can handle this situation. I am using pair of IOLOCK and IOUNLOCK function but it results in system hang Thanking in advance Santosh k -- Life is Small and Lots of Things to Learn Lets Share the Knowledge _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0

Carbon File Manger VFS caching
by Brian Bergstrand 12 Aug '05

12 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Thanks. Brian Bergstrand <http://www.bergstrand.org/brian/> PGP Key ID: 0xB6C7B6A2 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) iD8DBQFC/TLWedHYW7bHtqIRAp0cAJ9tOGlaG+YFlv1y9VT3bNkAwDytSACg31g7 9UDCmtQTsVh491BPiNHiaPw= =icb8 -----END PGP SIGNATURE----- _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… We've had problems with the Carbon File Mgr caching dir entries for too long in Panther and Jaguar and the problem persists in Tiger. Sometimes it takes more than a minute to refresh and sometimes it seems to never refresh. Our timestamps are cached in the VFS for 10 seconds, and then we hit the network again. Is there anything else that we can do to get a refresh quicker? We set FAST_STATFS in the vol attributes, but it didn't make a difference. So far, only quitting the Finder or running the following script can force a refresh and neither are very useful in day to day use by customers. tell application "Finder" try set mytarget to POSIX path of (target of front window as Unicode text) do shell script "/bin/ls -l " & quoted form of mytarget update target of front window end try end tell This email sent to site_archiver(a)lists.apple.com

1 0

Re: IOSCSIArchitectureModelFamily-140.1.7
by Herb Petschauer 12 Aug '05

12 Aug '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(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:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=jifLphUYEnQzS/zsSxOwDccC7ImaXDQLD4aPLYy8pIWdWZ0BnW56NOFacTKvmssbbQmXu3Dqutwqeepr0yqVVT+NsjZA213mjN+uvJMx0MiSjvsuBnHqkv5RAxcR64f/aisZaiUsMQfKjpF4SpBFGiUMd9/85fxlngPZWvfbess= Ok, it's not just me: "Sorry, the IOSCSIArchitectureModelFamily sources for Tiger are unavailable at this time." Sorry for the wasted bandwidth. Cheers, H. On 8/12/05, Herb Petschauer <bogvardi(a)gmail.com> wrote: > I haven't seen or been able to find a question on this, apologies if > it has been addressed already. > > I'm looking for IOSCSIArchitectureModelFamily-140.1.7, I get the > "Download Unavailable" message when I try and download it from > "Darwin". It's difficult to build IOSCSIParallelController without it > ;-) > > Is it just me? > > Cheers, > > Thanks! > H. > _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.… This email sent to site_archiver(a)lists.apple.com

1 0