January 2005 - Darwin-kernel

ntp_adjtime() syscall problems
by Sean Peisert 07 Jan '05

07 Jan '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:reply-to:to:subject:mime-version:content-type:content-transfer-encoding; b=skeWTT7jKxSrkexwHt9viZtSRWOpMrVflf8G5awLYLN6EpZlJ0akdTVod/8RoUkBiiptisDTpF0B5UUamrCxaJklrkmUN1cnn6VGqZ4CqM5u61G3jlFxx4xnjSODcgyf9hw3p2EYJbknEUQFXVsgcAX25z0kp2b4gc5FxR8NV3g= I am attempting to add the ntp_adjtime() system call present on FreeBSD systems to Darwin. It is needed for an Internet2 network performance measurement project where microsecond-level timing accuracy (or better) is required from NTP. I already have CDMA timing clocks hooked in through the serial ports working properly. To do this, I grabbed the timex.h and kern_ntptime.c files from the FreeBSD distribution, and added ntp_adjtime() into the xnu kernel as system call #369, and compiled successfully. Although ntpq -p reports the times succesfully with a high degree of accuracy, when I attempted to use the owamp and bwctl applications (from Internet2), which rely on the ntp_adjtime() call that I added, to query ntp, I received UNSYNC errors, which is both bizarre and inaccurate. To explore further, I built the ntptime application from Apple's ntp-16 distribution by commenting out the ntp_gettime() calls and replacing the ntp_adjtime(x) calls with syscall(369,x). Unfortunately, ntptime only returns errors, as follows: /tmp/ntp/Build/util root# ./ntptime ntp_gettime() returns code 5 (ERROR) time 83a27e80.b580f000 Thu, Dec 25 1969 14:21:52.708, (.708999), maximum error 3259744661 us, estimated error 2273678227 us, TAI offset -1035222635 ntp_adjtime() returns code 5 (ERROR) modes 0x0 (), offset 0.000 us, frequency 0.000 ppm, interval 1 s, maximum error 500000 us, estimated error 500000 us, status 0x40 (UNSYNC), time constant 0, precision 0.000 us, tolerance 496 ppm, I'm assuming that adding this system call and making it work successfully isn't as simple as what we did. There are problem status/control codes to set in the timex.h file that aren't set properly for Mac OS X 10.3.x Server, as they are for FreeBSD. Alternatively, perhaps there's something else going on between NTP and the kernel? Does anyone have any thoughts on how to fix the situation? Any changes I can make to timex.h or kern_ntptime.c? Or anything else? Many thanks, Sean -- Sean Peisert http://www.sdsc.edu/~peisert/ Fellow, San Diego Supercomputer Center University of California, San Diego _______________________________________________ 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

VFS-Lock patch on MAC OS X
by Yogesh P 05 Jan '05

05 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hi Folks, I am working on the VFS-Lock patch on linux which gives the functionality for locking and unlocking of filesystem. Is there any patch available on MAC for locking and unlocking of filesystem for a fraction of time????OR is the darwinkernel has VFS-Lock functionality built in???? Any pointers,reply certainly helps. Thanks, YOGESH PAHILWAN (SOFTWARE ENGINNER) S P Software Technologies (I) Pvt. Ltd. (An ISO 9001:2000 Company) Pune, Maharashtra, India - 411 004. Tel : 91-20-4006154, 4006983, 4007503. Fax : 91-20-4006153. Email : pahilwan.yogesh(a)spsoftindia.com Visit us at : www.spsoftindia.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

1 0

Top layering of mass storage driver stack
by Yogesh P 05 Jan '05

05 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hi Folks, I am working on the AppleRAID kext. I know that the AppleRAID kext is placed at the top layer in the mass storage driver stack means at device service layer.Its layering is above block storage driver and below BSD layer. Suppose, i created the logical partitions or raid partitions using disk-utility which underneath use AppleRAID for the operations. If i want to perform the read/write operations on these partitions,does i need to write partition scheme driver or filter scheme driver??? And at what layer my driver will be reside???Is it reside above the layer of AppleRAID kext or is it placed at the same layer where AppleRAID is???? Any pointers,reply certainly helps. Thanks, YOGESH PAHILWAN (SOFTWARE ENGINNER) S P Software Technologies (I) Pvt. Ltd. (An ISO 9001:2000 Company) Pune, Maharashtra, India - 411 004. Tel : 91-20-4006154, 4006983, 4007503. Fax : 91-20-4006153. Email : pahilwan.yogesh(a)spsoftindia.com Visit us at : www.spsoftindia.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

1 0

Re: UDF 2.0 Support, Hitachi DZMV550A Video Camera
by David Gatwood 03 Jan '05

03 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com near-complete rewrite. There has been talk about writing a "driver driver": an IOkit driver that is really just a wrapper around a *BSD or Linux driver, allowing for very simple porting. Nobody has gotten around to it, and I doubt they will because I think its a bit harder to do than it sounds. It isn't necessarily that much harder, but each family has specific design decisions that require specific wrapper designs. In some cases, the abstraction layering differs enough that it would be difficult. In other cases, not so much. Your mileage may vary. There's a driver porting guide in the Darwin section of the developer site that might prove helpful if you decide to wrap a driver. It won't help with family-specific issues, but it at least should help with some of the more common general differences between the BSD driver model and the I/O Kit. It includes some code snippets that should be helpful (I hope) when wrapping drivers from other OSes. It doesn't have a full wrapper, though, as that would vary significantly from family to family. David _______________________________________________ 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.… On Dec 31, 2004, at 12:04 PM, John Davidorff Pell <johnpell(a)mac.com> wrote: This email sent to site_archiver(a)lists.apple.com

1 0

Re: Interception of write system call at block level and not at FS level
by Garth Cummings 03 Jan '05

03 Jan '05

site_archiver(a)lists.apple.com Delivered-To: darwin-kernel(a)lists.apple.com Hello Yogesh, On Dec 27, 2004, at 10:12 PM, Yogesh P wrote: HI, I would like to implement a OSX file system call interception such as write call at block level and not at file system level(below FS level). Regards, --gc __________________________________________________________________ Garth Cummings Apple Developer Technical Support gcummings(a)apple.com http://developer.apple.com/technicalsupport _______________________________________________ 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 It sounds like a filter scheme would meet your requirements. Please see <http://developer.apple.com/documentation/DeviceDrivers/Conceptual/ MassStorage/07_Media_Example/chapter_7_section_1.html>. The sample code in the document is published separately here: <http://developer.apple.com/samplecode/SampleFilterScheme/ SampleFilterScheme.html> smime.p7s

1 0