February 2002 - Darwin-kernel

Re: A few questions...
by Emiel Kollof 26 Feb '02

26 Feb '02

Louis Gerbarg heeft op dinsdag 26 februari 2002 om 21:23 het volgende geschreven: [snip] No.Top of tree is not something you want, it may not work, or it may have problems. You want to pull a tagged release. Also, newer kernels may have dependencies on other packages, like Libc, etc. Simply pulling hte newest kernel and trying to run it on a release of Darwin or OS X may have signficant issues. I noticed... It doesn't even buld correctly (On a Mac OS X 10.1.3 box). So I'll guess I'll just patch myself down to Apple-201-15 or something then. (Not that I mind running bleeding edge. I run NetBSD and FreeBSD current as well). Thanks for the info anyway. On a side note: is ipv6 useable yet? I notice there is no userspace support for ipv6. Has KAME released anything yet that can reliably be built on OSX 10.1.3? Cheers, Emiel Kollof _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: A few questions...
by Louis Gerbarg 26 Feb '02

26 Feb '02

On Tuesday, February 26, 2002, at 11:28 AM, Emiel Kollof wrote: Louis Gerbarg heeft op dinsdag 26 februari 2002 om 19:52 het volgende geschreven: [snip] Apple-222 is different than Apple-201-5. The deal is when new kernels are submitted internally they are generally given a new montonically increasing number. When the source for those is checked into Darwin with the tag they are checked in under being Apple-nnn where nnn is the version of the project. The notable difference between the kernels you listed is that Apple-201 was the kernel that shipped with Mac OS X 10.1 and Darwin 1.4.1. After that stuff for software updates was done on the Apple-201-xx branch, and more work not targetted at the software update kept having increasing numbers. There is no -STABLE or -CURRENT as in FreeBSD, Interesting... So if I want to track xnu (and other modules) I just have to follow the default branch (what you get if you don't supply a branch with -r). No.Top of tree is not something you want, it may not work, or it may have problems. You want to pull a tagged release. Also, newer kernels may have dependencies on other packages, like Libc, etc. Simply pulling hte newest kernel and trying to run it on a release of Darwin or OS X may have signficant issues. Louis _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: A few questions...
by Emiel Kollof 26 Feb '02

26 Feb '02

Louis Gerbarg heeft op dinsdag 26 februari 2002 om 19:52 het volgende geschreven: [snip] Apple-222 is different than Apple-201-5. The deal is when new kernels are submitted internally they are generally given a new montonically increasing number. When the source for those is checked into Darwin with the tag they are checked in under being Apple-nnn where nnn is the version of the project. The notable difference between the kernels you listed is that Apple-201 was the kernel that shipped with Mac OS X 10.1 and Darwin 1.4.1. After that stuff for software updates was done on the Apple-201-xx branch, and more work not targetted at the software update kept having increasing numbers. There is no -STABLE or -CURRENT as in FreeBSD, Interesting... So if I want to track xnu (and other modules) I just have to follow the default branch (what you get if you don't supply a branch with -r). 3. Am I on the right mailing list for questions like these? Or should I move this to darwin-development? This kind of questions is probably appropriate for either list, not sure which one it is more appropriate on. Oh well, at least I got some answers :) Thanks for your time. I will not return to my regular programming and submit build errors, bugs or other stuff that ticks me :) Cheers, Emiel _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: A few questions...
by Louis Gerbarg 26 Feb '02

26 Feb '02

On Monday, February 25, 2002, at 01:46 PM, Emiel Kollof wrote: Hi people, I am new to this list, and I have some questions: 1. Where is the kernel FAQ with answers to questions like these? I can't seem to find it anywhere, so I'll ask here. There are some general Darwin faqs with some kernel info at www.darwinfo.org. but I do not think there is any "Kernel FAQ" (someone please chime in if I am wrong). 2. Which branch is the one to track xnu with CVS? There are many, and the branch numbering is inconsistent (ie. Apple-222 is a lower kernel version than say Apple-201-5). Can someone point me to a document that explains which branch is the one to track? What is the development style like? Like FreeBSD's? If yes, which is the equivalent of CURRENT, and which STABLE? Apple-222 is different than Apple-201-5. The deal is when new kernels are submitted internally they are generally given a new montonically increasing number. When the source for those is checked into Darwin with the tag they are checked in under being Apple-nnn where nnn is the version of the project. The notable difference between the kernels you listed is that Apple-201 was the kernel that shipped with Mac OS X 10.1 and Darwin 1.4.1. After that stuff for software updates was done on the Apple-201-xx branch, and more work not targetted at the software update kept having increasing numbers. There is no -STABLE or -CURRENT as in FreeBSD, 3. Am I on the right mailing list for questions like these? Or should I move this to darwin-development? This kind of questions is probably appropriate for either list, not sure which one it is more appropriate on. Louis _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

serial console
by Andrew Gallatin 26 Feb '02

26 Feb '02

I've gone out and purchased a "G4 port" serial port for my development g4 box. I'm able to connect to it via my console server and talk to the Openfirmware. However, as soon as I boot OS X, it switches to a graphics console: 1 > boot load-size=1b8b2 adler32=a92ee041 parsing <CHRP-BOOT> evaluating <BOOT-SCRIPT> loading XCOFF tsize=18000 dsize=3000 bsize=99000 entry=1c13594 SECTIONS: .text 01c00000 01c00000 00018000 000000d4 .data 01c18000 01c18000 00003000 000180d4 .bss 01c1b000 01c1b000 00099000 00000000 loading .text, done.. loading .data, done.. clearing .bss, done.. <kernel messages now on graphics head> How do I force darwin to use a serial console? Thanks, Drew _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

A few questions...
by Emiel Kollof 25 Feb '02

25 Feb '02

Hi people, I am new to this list, and I have some questions: 1. Where is the kernel FAQ with answers to questions like these? I can't seem to find it anywhere, so I'll ask here. 2. Which branch is the one to track xnu with CVS? There are many, and the branch numbering is inconsistent (ie. Apple-222 is a lower kernel version than say Apple-201-5). Can someone point me to a document that explains which branch is the one to track? What is the development style like? Like FreeBSD's? If yes, which is the equivalent of CURRENT, and which STABLE? 3. Am I on the right mailing list for questions like these? Or should I move this to darwin-development? Cheers (and thanks for your time, I hope I haven't wasted any), Emiel Kollof _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: LXR for darwin?
by Justin C. Walker 21 Feb '02

21 Feb '02

On Thursday, February 21, 2002, at 09:57 AM, Sujal Shah wrote: Is the darwin source code in LXR anywhere (or any other cross referencer)? Would Apple object if I were to put one up (assuming I have enough web space to do so)? I don't believe that there is any externally available LXR setup for darwin. You should contact Apple directly about doing this, since there may be licensing issues. Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large * Institute for General Semantics | Men are from Earth. | Women are from Earth. | Deal with it. *--------------------------------------*-------------------------------* _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Unresolved Kernel Trap
by Robert Owen 21 Feb '02

21 Feb '02

I originally posted this to the server discussion list, however someone there suggested I post it here. So.... I recently updated our server to Mac OS X 10.1.2 Server (I had applied all other patches as they came out, including the 10.1.2 client update). Upon restart, the machine froze with an unresolved Kernel Trap. This error seems to happen at various times during the startup process and sometimes not even at all. However, if I startup without the ethernet cable in, it starts up fine. I can then stick the cable back in, open the system prefs, turn appletalk off and then back on and I have a functional server again (as long as I don't have to restart it.) Restarting the computer also causes a panic. I have also noticed that I am unable to startup from the OS X 10.1 Server CD (to check the drive for errors or reinstall). Again the server hangs with an unresolved Kernel Trap, however it never makes it past the happy mac screen and I only see the error in verbose mode. I can boot with the CD in if I don't have the ethernet cable in. I have the error messages as they are generated in various forms as screen shots (gotta love digital cameras). If anyone thinks they might be able to help me work this out, I can provide the information to them. I am interested in trying to solve this, however if anyone thinks it is not worth the time and a reinstall is a better idea, I am open to that also. Thanks in advance. Robert Owen -- -------------------------------------------------------------------------------------------- Robert M. Owen tel: (608) 262-9234 Kiessling Group email: rowen(a)chem.wisc.edu http://www.chem.wisc.edu/~kiessling Department of Chemistry University of Wisconsin-Madison 1101 University Ave. Madison, WI 53706 _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: IOPanic ans assert not working ?
by Umesh Vaishamapayan 10 Feb '02

10 Feb '02

On Tuesday, February 5, 2002, at 05:46 PM, Justin C. Walker wrote: Ahh. I see.. I'm used to alphas where you can generally switch back to the SRM console firmware for I/O if you're in a bad state.. On the macs, they quiesce the OF, which means that you can't use OF drivers in the kernel. I can only think of one OS that doesn't do this (NetBSD). It's just a lot of effort to keep OF around. And, of course, Solaris, where OF was first perpetrated on an unsuspecting world. ... and NeXTSTEP on sparc. Came in handy when network driver did not co-operate [and while debugging KDP]. One of the engineers wrote Forth code to read mach_kernel symbol table. That enabled symbolic backtraces in OF! --Umesh _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0

Re: NFS limited to 4K reads/writes
by Umesh Vaishamapayan 10 Feb '02

10 Feb '02

On Thursday, January 31, 2002, at 08:52 AM, Andrew Gallatin wrote: I was playing around with NFS. Darwin seems to perform well as a server, but its performance leaves a lot to be desired as a client because it appears to crippled by a 4k read and write size limit. (verified w/tcpdump). I can reduce the size with mount_nfs, but I cannot increase it over 4K bsd/nfs/nfs_bio.c: 108 int 109 nfs_bioread(vp, uio, ioflag, cred, getpages) 110 register struct vnode *vp; ... 138 /*due to getblk/vm interractions, use vm page size or less values */ 139 biosize = min(vp->v_mount->mnt_stat.f_iosize, PAGE_SIZE); ... pagein/pageout paths do not have this limitation... By poor performnace, I mean that Darwin reads at 12MB/sec over a 2Gb/sec link. Tru64 and FreeBSD can both read at 50MB/sec. (this is from a file warm in the server's cache; no disk I/O). Darwin performs just as well as the competition when its the server... Any chance of this 4K limit being lifted? In order to enable larger read()/writes(), vnode_pageout() and following NFS functions will needs some tweaks... bsd/nfs/nfs_bio.c nfs_bioread 139 biosize = min(vp->v_mount->mnt_stat.f_iosize, PAGE_SIZE); bsd/nfs/nfs_bio.c nfs_write 543 biosize = min(vp->v_mount->mnt_stat.f_iosize, PAGE_SIZE); bsd/nfs/nfs_bio.c nfs_write 640 if (!ISSET(bp->b_flags, B_CACHE) && n < PAGE_SIZE) { bsd/nfs/nfs_bio.c nfs_write 645 auio.uio_resid = PAGE_SIZE; bsd/nfs/nfs_bio.c nfs_write 650 iov.iov_len = PAGE_SIZE; bsd/nfs/nfs_bio.c nfs_write 660 bp->b_validend = PAGE_SIZE - auio.uio_resid; bsd/nfs/nfs_bio.c nfs_write 663 PAGE_SIZE); bsd/nfs/nfs_bio.c nfs_getcacheblk 849 int biosize = min(vp->v_mount->mnt_stat.f_iosize, PAGE_SIZE); bsd/nfs/nfs_vfsops.c nfs_iosize 223 if (iosize < PAGE_SIZE) bsd/nfs/nfs_vfsops.c nfs_iosize 224 iosize = PAGE_SIZE; bsd/nfs/nfs_vnops.c nfs_setattr 974 PAGE_SIZE); bsd/nfs/nfs_vnops.c nfs_pageout 4402 lbn = f_offset / PAGE_SIZE; bsd/nfs/nfs_vnops.c nfs_pageout 4404 for (iosize = size; iosize > 0; iosize -= PAGE_SIZE, lbn++) { bsd/nfs/nfs_vnops.c nfs_pageout 4450 pgsize = (iosize + (PAGE_SIZE - 1)) & ~PAGE_MASK; bsd/nfs/nfs_vnops.c nfs_blktooff 4595 biosize = min(vp->v_mount->mnt_stat.f_iosize, PAGE_SIZE); bsd/nfs/nfs_vnops.c nfs_offtoblk 4613 biosize = min(vp->v_mount->mnt_stat.f_iosize, PAGE_SIZE); --Umesh -- Umesh Vaishampayan Mac OS X kernel group _______________________________________________ darwin-kernel mailing list | darwin-kernel(a)lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.

1 0