March 2003 - Darwin-kernel - lists.apple.com

Re: darwin-kernel digest, Vol 2 #243 - 3 msgs
by William Kucharski 10 Mar '03

10 Mar '03

>In the last case [ if (ISSET(vp->v_flag, VUINIT)) ], where the thread >sleeps waiting for the VNODE to finish init, the interlock is dropped >before waiting, but after returning from the wait it jumps immediately >to retry: w/o checking flags for LK_INTERLOCK. If LK_INTERLOCK is set, >then the interlock won't be acquired at the beginning of retry:. Maybe >I'm missing something, but before going to retry, shouldn't >LK_INTERLOCK be cleared or if set then then re-acquire the lock before >the jump? What am I missing? The first lines of the routine are: retry: /* * If the vnode is in the process of being cleaned out for * another use, we wait for the cleaning to finish and then * return failure. Cleaning is determined by checking that * the VXLOCK flag is set. */ if ((flags & LK_INTERLOCK) == 0) simple_lock(&vp->v_interlock); so when the VUINIT leg calls simple_unlock and sleeps, when it wakes and does a goto retry the lock will be reacquired if LK_INTERLOCK is still 0. LK_INTERLOCK is not set by the simple_lock() call, or by any intervening code in vget(), so no flags are set that need to be unset; the release of v_interlock is sufficient. William Kucharski kucharsk(a)mac.com _______________________________________________ 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: error initializing loopback device in 10.2.4
by Felix Kronlage 10 Mar '03

10 Mar '03

On Sun, Mar 09, 2003 at 05:06:29PM -0500, Vasisht Tadigotla wrote: > Is it possible to get the 6.4 sources from cvs ? afaik Apple has not released the 6.4 (= 10.2.4) source yet. -fkr -- gpg-fingerprint: 076E 1E87 3E05 1C7F B1A0 8A48 0D31 9BD3 D9AC 74D0 |Felix Kronlage | whois -h whois.ripe.net FKR-RIPE | |http://www.hazardous.org/fkr/ | fkr(a){grummel.net|opendarwin.org} | [demime 0.98b removed an attachment of type application/pgp-signature] _______________________________________________ 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: Locking bug in vget()?
by Quinn 10 Mar '03

10 Mar '03

At 1:21 -0600 9/3/03, Brian Bergstrand wrote: I think I may have found a bug in vget(). I took a quick look at this and it looks like a problem to me too. Please file a bug so that one of our kernel engineers can look at this and makes a final determination. <http://developer.apple.com/bugreporter/> S+E -- Quinn "The Eskimo!" <http://www.apple.com/developer/> Apple Developer Technical Support * Networking, Communications, Hardware _______________________________________________ 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

Happy child (w/10.2), gets SIGABRT (w/10.1)
by Lance Drake 10 Mar '03

10 Mar '03

It is only after having spent my entire weekend poking and prodding and testing and reading and looking and loading and thinking and trying... that this letter has been posted to this list. So please, don't think I'm being flip and just tossing my trivial problems into your collective laps. Manager's Nutshell Explanation: ------------------------------- The (simple) tool that builds with PB 2.0.1 on MacOSX 10.2.4 (and works perfectly) does not work when built with PB 2.0 on MacOSX 10.1. After a call to wait4(ch_pid...), the parent, examining WIFEXITED, reports the child process has been quit via SIGABRT. (100% repeatable) Discussion: ----------- In my tool, main() performs a fork(), the parent is supposed to quit and leave the child remaining and running. This is exactly what happens with the 10.2.4 implementation. When my manager asked for 10.1 support, it seemed like a simple enough thing - but noooooo... I have loaded a separate disk with OS 10.1 and installed the 2.0 PB tools, re-built the project and the formerly, completely reliable code sputters to a total standstill. If I skip the fork and simply execute, all of my code runs with no problems or exceptions - it works fine! But, as soon as the fork takes place, the child process dies a horrible death... and I have reached the point of admitting I am clueless. For sure, processes and fork() are nothing new - certainly they work with 10.1 but scrounging the Q&A, TechNotes, Docs, Mach literature, UNIX books, Google and the archives of this newsgroup give no clue as to what is happening in my case. I am not even sure where the SIGABRT is coming from - and certainly not WHY is is coming. Wanna make somebody deliriously happy? Share some of your secret-inside-coolguy-incantations and help put this forkin' brickwall behind me. Thanks! Lance Drake _______________________________________________ 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: error initializing loopback device in 10.2.4
by Vasisht Tadigotla 09 Mar '03

09 Mar '03

On Fri, 7 Mar 2003, Justin Walker wrote: > > One possible explanation is that this is a timing problem: the IPv6 > initialization of 'lo0' is occurring before the system has set up the > mbuf pool. I have seen this before, in early versions of Darwin, but I > thought that had been fixed. it seems to be a problem only with Darwin 6.4, I compiled Darwin 6.3 from cvs and I didn't have the problem with the loopback device when I use the 6.3 kernel. The sources for 6.4 doesn't seem to have made it to the cvs yet. I haven't heard anyone else having the same problem, something weird happening with my machine perhaps ?? Is it possible to get the 6.4 sources from cvs ? I'd be grateful If anyone can point out how to. Thanks, Vasisht --------------------------------------------------------------------- Life is complex: it has real and imaginary components _______________________________________________ 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

Locking bug in vget()?
by Brian Bergstrand 09 Mar '03

09 Mar '03

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I think I may have found a bug in vget(). Here is the code from vfs_subr.c:1.1.1.6 int vget(vp, flags, p) struct vnode *vp; int flags; struct proc *p; { int error = 0; retry: /* * If the vnode is in the process of being cleaned out for * another use, we wait for the cleaning to finish and then * return failure. Cleaning is determined by checking that * the VXLOCK flag is set. */ if ((flags & LK_INTERLOCK) == 0) simple_lock(&vp->v_interlock); if ((vp->v_flag & VXLOCK) || (vp->v_flag & VORECLAIM)) { vp->v_flag |= VXWANT; simple_unlock(&vp->v_interlock); (void)tsleep((caddr_t)vp, PINOD, "vget", 0); return (ENOENT); } /* * vnode is being terminated. * wait for vnode_pager_no_senders() to clear VTERMINATE */ if (ISSET(vp->v_flag, VTERMINATE)) { SET(vp->v_flag, VTERMWANT); simple_unlock(&vp->v_interlock); (void)tsleep((caddr_t)&vp->v_ubcinfo, PINOD, "vclean", 0); return (ENOENT); } /* * if the vnode is being initialized, * wait for it to finish initialization */ if (ISSET(vp->v_flag, VUINIT)) { if (ISSET(vp->v_flag, VUINIT)) { SET(vp->v_flag, VUWANT); simple_unlock(&vp->v_interlock); (void) tsleep((caddr_t)vp, PINOD, "vget2", 0); goto retry; } } ... In the last case [ if (ISSET(vp->v_flag, VUINIT)) ], where the thread sleeps waiting for the VNODE to finish init, the interlock is dropped before waiting, but after returning from the wait it jumps immediately to retry: w/o checking flags for LK_INTERLOCK. If LK_INTERLOCK is set, then the interlock won't be acquired at the beginning of retry:. Maybe I'm missing something, but before going to retry, shouldn't LK_INTERLOCK be cleared or if set then then re-acquire the lock before the jump? Something like this: /* * if the vnode is being initialized, * wait for it to finish initialization */ if (ISSET(vp->v_flag, VUINIT)) { if (ISSET(vp->v_flag, VUINIT)) { SET(vp->v_flag, VUWANT); simple_unlock(&vp->v_interlock); (void) tsleep((caddr_t)vp, PINOD, "vget2", 0); flags &= ~LK_INTERLOCK; goto retry; } } Brian Bergstrand <http://www.classicalguitar.net/brian/> PGP Key ID: 0xB6C7B6A2 We are not retreating - we are advancing in another direction. - General Douglas MacArthur -----BEGIN PGP SIGNATURE----- Version: PGP 8.0 iQA/AwUBPmrdjHnR2Fu2x7aiEQK3DwCg4ysw54MxU0zrNpWMn7lQy8a2EmQAn3qO dBTnYTyYAMdbpmiHNKY+t8Ao =Qm9E -----END PGP SIGNATURE----- _______________________________________________ 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: Remote Kernel debugging
by Paul Ripke 08 Mar '03

08 Mar '03

On Tuesday, Mar 4, 2003, at 21:29 Australia/Sydney, Pev wrote: Hi chaps! I'm looking at setting up a remote debug box for my dev machine (G4 PM) but I'm wondering what debug configurations are known to work. I don't have a second mac, so thats not a possibility for me, but I do have a few old PC's around. Has anyone had success using an x86 darwin box for remotely debugging a mac running darwin? I'm not sure if this is likly to be a big nono or not...! "real" gdb supports cross-target debugging, so I figured I'd have a quick look. The Apple changes don't appear to have made it into the latest gdb release, so I grabbed the gdb tarball from Apple's site, and tried building under i386 NetBSD for a powerpc-apple-macos10 target. Configure ran fine, and the build seemed to compile much of gdb, but broke down on macosx files with header dependencies (mostly mach related). I tried to give it what it wanted, but there's too many differences and would take more time than I was prepared to spend to see if I could get it past the compile stage. My guess is you may have far more luck on x86 Darwin, I'm guessing it may even build clean. No bets as to whether it would work, though. As always, try it and see! Cheers, -- Paul Ripke Unix/OpenVMS/TSM/DBA 101 reasons why you can't find your Sysadmin: 68: It's 9AM. He/She is not working that late. -- Koos van den Hout _______________________________________________ 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: error initializing loopback device in 10.2.4
by Vasisht Tadigotla 08 Mar '03

08 Mar '03

> No, I don't think that is the problem. Where do you see the error > message? almost at the very beginning of bootup, right after the initialization of the PCMCIA slot. > One possible explanation is that this is a timing problem: the IPv6 > initialization of 'lo0' is occurring before the system has set up the > mbuf pool. I have seen this before, in early versions of Darwin, but I > thought that had been fixed. > > What kind of machine are you using, and when did this start happening? It's a Ti Book 667, this started happening round about the time I upgraded to 10.2.4. I think your explanation might explain why when I force shutdown things are back to normal. thanks, Vasisht --------------------------------------------------------------------- Life is complex: it has real and imaginary components _______________________________________________ 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: error initializing loopback device in 10.2.4
by Justin Walker 07 Mar '03

07 Mar '03

Hi, On Thursday, Mar 6, 2003, at 17:33 US/Pacific, Vasisht Tadigotla wrote: hi all, i'm getting this error initializing the loopback device for ipv6. the error message is in6_ifattach_loopback: failed to configure the loopback address on lo0 (errno=55). errno=55 is ENOBUFS, but i can get the loopback device up manually doing sudo ifconfig lo0 inet6 ::1 up also if i cause a forced shutdown of the system, the loopback device is setup normally. should i increase the number of mbufs for my system, it's a ti book 667. No, I don't think that is the problem. Where do you see the error message? One possible explanation is that this is a timing problem: the IPv6 initialization of 'lo0' is occurring before the system has set up the mbuf pool. I have seen this before, in early versions of Darwin, but I thought that had been fixed. What kind of machine are you using, and when did this start happening? Regards, Justin -- Justin C. Walker, Curmudgeon-At-Large * Institute for General Semantics | It's not whether you win or lose... | It's whether *I* win or lose. *--------------------------------------*-------------------------------* _______________________________________________ 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

error initializing loopback device in 10.2.4
by Vasisht Tadigotla 07 Mar '03

07 Mar '03

hi all, i'm getting this error initializing the loopback device for ipv6. the error message is in6_ifattach_loopback: failed to configure the loopback address on lo0 (errno=55). errno=55 is ENOBUFS, but i can get the loopback device up manually doing sudo ifconfig lo0 inet6 ::1 up also if i cause a forced shutdown of the system, the loopback device is setup normally. should i increase the number of mbufs for my system, it's a ti book 667. the output of netstat -mb in both cases when loopback's been configured normally and when it fails are the same : 69 mbufs in use: 66 mbufs allocated to data 2 mbufs allocated to socket names and addresses 1 mbufs allocated to Appletalk data blocks 79/202 mbuf clusters in use 421 Kbytes allocated to network (41% in use) 0 requests for memory denied 0 requests for memory delayed 0 calls to protocol drain routines does anyone else have the same problem ? or know what's happening here ? thanks, vasisht --------------------------------------------------------------------- Life is complex: it has real and imaginary components _______________________________________________ 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