Re: smbfs breakage in 4.5-stable
On Sat, 2002-02-02 at 16:04, Peter C. Lai wrote: OK. Here's what I did wrong (and after fixing this, it worked): in /usr/src/sys/i386/conf/MACHNAME built from scratch, I had: OPTIONS NETSMB OPTIONS NETSMBCRYPTO OPTIONS LIBMCHAIN OPTIONS LIBICONV however, in the filesystems section, i had omitted OPTIONS SMBFS apparently, if I explicitly add net[smb|crypto], libmchain, and libiconv, I need to add in smbfs also. (I didn't realize this, since SMBFS is lumped in with all the other filesystems, which don't depend on anything, and can be safely modularized without being declared in the configuration file). Adding SMBFS back in to the kernel config fixed the smbfs problem. Now, can I remove net[smb|crypto], libmchain, libiconv, and smbfs, and have everything working as modules? Implicit modularization still confuses me (unlike in linux, where you explicitly state static,module,or don't-compile). I have _nothing_ relating to smbfs in my kernel, and mount_smbfs works just using klds. The only place I've seen a problem with just using klds is on SMP machines. In those cases, I had to statically include the smbfs stuff in the kernel. Joe On Sat, Feb 02, 2002 at 01:29:58AM -0500, Joe Clarke wrote: On Sat, 2002-02-02 at 01:09, Peter C. Lai wrote: After 4.5 was released, i cvsup'd a complete fresh set of sources (e.g. deleting /usr/src/* and /usr/obj/*), made world, installed world and a new kernel, and now mount_smbfs errors out the first time with 'exec format' and if you try it again, it says 'file exists'. kldstat shows the smbfs.ko not loaded the first time it fails but it does load the second time, and remains loaded. There is no error with the module loading. The smbfs kernel module was modified in the 4.5-PRERELEASE track so to depend on the libiconv and libmchain modules. Before that, doing a mount_smbfs resulted in an exec format error with a message in /var/log/messages indicated a symbol in smbfs.ko was undefined. You might want to verify that all of those modules are up-to-date. I just tried this on a recent 4.5-RC machine with no smbfs-related options in the kernel, and it worked like a champ. After perusing the mailing lists, I realized the problem could be caused by the smbfs port which i didn't remove prior to make and install world. I pkg_delete'd it, during which it noted that /sbin/mount_smbfs and /modules/smbfs.ko had been modified. Thus, the installworld did overwrite those two files into the 4.5 versions. Make sure that when you rebuild world, you build all the kernel modules as well. When you mount a smb share, you should have the following modules loaded: 51 0xc3417000 1a000smbfs.ko 61 0xc3432000 3000 libiconv.ko 71 0xc3436000 3000 libmchain.ko Joe on bootup, netsmb_dev loads fine. Another note: I installed the same world onto another machine running a completely new install of 4.5-rc3 via samba export, and now that machine's mount_smbfs is broken too. any help would be appreciated Please cc responses to [EMAIL PROTECTED] since i don't subscribe to either -questions or -stable -- Peter C. Lai University of Connecticut Dept. of Residential Life | Programmer Dept. of Molecular and Cell Biology | Undergraduate Research Assistant http://cowbert.2y.net/ 860.427.4542 (Room) 860.486.1899 (Lab) 203.206.3784 (Cellphone) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message -- Peter C. Lai University of Connecticut Dept. of Residential Life | Programmer Dept. of Molecular and Cell Biology | Undergraduate Research Assistant http://cowbert.2y.net/ 860.427.4542 (Room) 860.486.1899 (Lab) 203.206.3784 (Cellphone) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: problems CVS updating from 4.3 to 4.4 and 4.5
On Sat, Feb 02, 2002 at 02:40:18PM -0500, Michael Scheidell wrote: On Fri, Feb 01, 2002 at 04:06:34PM -0500, Michael Scheidell wrote: I believe this was fixed in 4.4-STABLE. reason I tried 4.4 rel is that upgrading to 4.5 rel failed also (some symlink to one of the iso lang files) Well, thats not the same problem then :) It shouldn't be happening if you're following the standard upgrade procedure (it _will_ if you don't: it often bites me when I try and be clever and skip steps). Kris msg41086/pgp0.pgp Description: PGP signature
stable.freebsd.org AKA releng4.freebsd.org down for next 2 days
Most people don't mirror anything from this machine, but just for the few who do, please consider this a HEADS UP! The machine is being replaced by a much beefier and faster machine, courtesy of Qwest, and will be back up just as soon as its new incarnation is clearly doing everything the old one did. Now when I say down I also don't mean actually down 24/7, and I'll actually endevor to keep the service mostly up for those two days, I just want you all to know that I'll feel free to reboot it or take it off-line at any time and without notice during that period until the service is transitioned. Thanks for your patience. - Jordan To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: problems CVS updating from 4.3 to 4.4 and 4.5
Kris Kennaway wrote: On Sat, Feb 02, 2002 at 02:40:18PM -0500, Michael Scheidell wrote: On Fri, Feb 01, 2002 at 04:06:34PM -0500, Michael Scheidell wrote: I believe this was fixed in 4.4-STABLE. reason I tried 4.4 rel is that upgrading to 4.5 rel failed also (some symlink to one of the iso lang files) Well, thats not the same problem then :) It shouldn't be happening if you're following the standard upgrade procedure (it _will_ if you don't: it often bites me when I try and be clever and skip steps). I remember switching to shell scripts to do all of this when O'Brien updated the /binutils during the transition from 4.0 4.1. The consequence of mistyping a script name was immediately apparent where as, mistyping an option would be found at the end of a bad make session. The sequence of buildworld, build[install]kernel, installworld wasn't an option when the /binutils were updated. I kind of wonder if the same thing is going to happen when the /binutils are updated in ~2 months as was commented else where. Kent -- Kent Stewart Richland, WA mailto:[EMAIL PROTECTED] http://users.owt.com/kstewart/index.html FreeBSD News http://daily.daemonnews.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: Junior Annoying Hacker Task
On Sat, 2002-02-02 at 14:50, Pete French wrote: AFAIK, Netinfo on Mac OS X is implemented deep. That is, it overrules standard libc behaviour (like the resolver, fstab and other things), Yes it's userspace jim, but not as we know it :) This was certainly true on NeXT's - you needed special versions of most programs (e.g. sendmail, bind, login etc...) that were netinfo aware too. If its going to be done it needs doing very thoroughly and carefully as it replaces more of the /etc files with netinfo equivalents. On the other doesnt YP do somethign similar ? (I've never had to use YP, thought I have been on the receiving end of some of the consequences). And I, somehow, had failed to notice that there's no NSS on FreeBSD. *smacks self on head* Not quite so simple after all, I guess -- brandon s. allbery [os/2][linux][solaris][japh] [EMAIL PROTECTED] system administrator [WAY too many hats][EMAIL PROTECTED] electrical and computer engineeringKF8NH carnegie mellon university [better check the oblivious first -ke6sls] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: problems CVS updating from 4.3 to 4.4 and 4.5
On Sat, Feb 02, 2002 at 01:37:41PM -0800, Kent Stewart wrote: I remember switching to shell scripts to do all of this when O'Brien updated the /binutils during the transition from 4.0 4.1. The consequence of mistyping a script name was immediately apparent where as, mistyping an option would be found at the end of a bad make session. The sequence of buildworld, build[install]kernel, installworld wasn't an option when the /binutils were updated. I kind of wonder if the same thing is going to happen when the /binutils are updated in ~2 months as was commented else where. Hang on, I thought that was the exact reason for the introduction of buildkernel/installkernel and the point in time when it was introduced. Kris msg41090/pgp0.pgp Description: PGP signature
Re: problems CVS updating from 4.3 to 4.4 and 4.5
Kris Kennaway wrote: On Sat, Feb 02, 2002 at 01:37:41PM -0800, Kent Stewart wrote: I remember switching to shell scripts to do all of this when O'Brien updated the /binutils during the transition from 4.0 4.1. The consequence of mistyping a script name was immediately apparent where as, mistyping an option would be found at the end of a bad make session. The sequence of buildworld, build[install]kernel, installworld wasn't an option when the /binutils were updated. I kind of wonder if the same thing is going to happen when the /binutils are updated in ~2 months as was commented else where. Hang on, I thought that was the exact reason for the introduction of buildkernel/installkernel and the point in time when it was introduced. Ah! Bad phrasing on my part. Wasn't an option meant it was the only choice. There was a race to be the first one to use the new /binutils the night he made the changes and I was the first one bitten by the old way of doing things. I haven't found any reason to switch since :). Kent -- Kent Stewart Richland, WA mailto:[EMAIL PROTECTED] http://users.owt.com/kstewart/index.html FreeBSD News http://daily.daemonnews.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: mpd-netgraph problem.
David Gilbert writes: I'm using mpd-netgraph to attempt to connect an encrypted tunnel. It appears to connect (according to the messages), but the following is spit out for most packets I try to put down the tunnel: [vpn] LCP: rec'd Protocol Reject #1 link 0 (Opened) [vpn] LCP: protocol 0x0029 was rejected [vpn] LCP: rec'd Protocol Reject #2 link 0 (Opened) [vpn] LCP: protocol 0x00a1 was rejected (on the one end) [strikeppp] rec'd unexpected protocol 0x0029 on link -1, rejecting [strikeppp] rec'd unexpected protocol 0x00a1 on link -1, rejecting [strikeppp] rec'd unexpected protocol 0x0001 on link -1, rejecting (on the other) The second log also contains lines of the form: [strikeppp] rec'd proto 0xee53 on MP link! (ignoring) [strikeppp] rec'd proto 0xcc0d on MP link! (ignoring) ... any ideas? This is usually because one side is sending encrypted traffic that the other is thinking is not encrypted... i.e., it's a side-effect of a negotiation problem. I've just heard from another person with this problem. Check your logs for something like ``enable chap required for MPPE'' on one side. As a workaround, if you are doing CHAP in both directions, try turning it off in one direction. -Archie __ Archie Cobbs * Packet Design * http://www.packetdesign.com To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: Junior Annoying Hacker Task
And I, somehow, had failed to notice that there's no NSS on FreeBSD. *smacks self on head* Not quite so simple after all, I guess Is this NSS as in the Mozilla crypt stuff ? If so then what does it have to do with netinfo - or is OSX netinfo different to normal netinfo ? [did they break yet another thing ? :-(] BTW, this discussion should probably not be taking place on as many lists as it currently is I suspect. Wheres most appropiate for it ? -pcf. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: CVSUP Mirror brain damage
In article [EMAIL PROTECTED], Michael Sierchio [EMAIL PROTECTED] wrote: It appears that cvsup2.freebsd.org is long gone. How to we get it off the list of mirrors, and/or of course see to it that it no longer is aliased to a host which isn't a cvsup mirror? I removed cvsup2 from the list of mirrors over a month ago, but I forgot to have its DNS record deleted. I have asked our hostmaster to do that now. We'll have a replacement for cvsup2 soon, maybe even this week. John -- John Polstra John D. Polstra Co., Inc.Seattle, Washington USA Disappointment is a good sign of basic intelligence. -- Chögyam Trungpa To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: Junior Annoying Hacker Task
On Sat, Feb 02, 2002 at 10:05:34PM +, Pete French wrote: And I, somehow, had failed to notice that there's no NSS on FreeBSD. *smacks self on head* Not quite so simple after all, I guess Is this NSS as in the Mozilla crypt stuff ? If so then what does it have to do with netinfo - or is OSX netinfo different to normal netinfo ? [did they break yet another thing ? :-(] NSS = Name Service Switch on any System V UNIX. BTW, this discussion should probably not be taking place on as many lists as it currently is I suspect. Wheres most appropiate for it ? As far as I'm concerned you can take it from all the lists. -- Zvezdan Petkovic [EMAIL PROTECTED] http://www.cs.wm.edu/~zvezdan/ To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: [wm-user] wdm
Has anyone been able to get wdm-120 working with FreeBSD 4.5-STABLE
and windowmaker-0.80.0?
Yes, using it right now:
andy@omni[191]- pkg_info | grep wdm
wdm-1.20WINGs Display Manager; an xdm replacement
andy@omni[192]- pkg_info | grep windowmaker
windowmaker-0.80.0 GNUStep-compliant NeXTStep window manager clone
andy@omni[193]- uname -a
FreeBSD omni.geek4food.org 4.5-RC FreeBSD 4.5-RC #75: Sun Jan 20 01:49:17 PST
2002 [EMAIL PROTECTED]:/usr/src/sys/compile/tureg i386
One thing I found was that wdm-1.20 was being linked against
libPropList. I removed that as from what I see it only needs
to be linked against libWINGS. That removed this error message
in the xdm-error.log:
wdmLogin error: could not open domain file /root/GNUstep/Defaults/wdmLogin
for reading: no such file or directory
I am still getting:
Caught signal 11.
Server aborting
I slo get this message on the console:
sigreturn: eflags = 0x13207
Ahhh. Now I've seen this (or something very similar, couldn't swear to the
exact hex sigreturn) a couple of times myself, and I never found out quite
what caused it.
Searching the FreeBSD archives never turned up anything (which is why I just
cross-posted the reply, in case someone else goes looking).
However, I did find out that it was basically caused by something in the
system being out-of-date. Which, as I cvsup all the time and don't necessarily
re-build the ports, isn't too surprising...
In one instance, rebuilding WindowMaker, WDM all the ports that it depended
on sorted it out. Another time, I had to completely rebuild *all* the ports
(doing the above plus X and a whole bunch of other stuff didn't sort it out,
so I did a pkg_delete `pkg_info | awk '{print $1}'` and started again).
I can only assume that something was stale in the system, and this was causing
the problem.
Any help would be greatly appreciated.
Vance, 'portupgrade' is your friend, if you've not looked at it, I can
heartily recommend it. It rocks.
My advice would be to pick a good moment to CVSup the complete source tree,
ports and all, go through the regular 'buildworld', 'buildkernel',
'mergemaster' cycle, and, having freshly rebooted into a good, clean, running
system, install 'portupgrade' if you don't have it already, fix up your ports
database and rebuild the whole lot.
At the end, the whole system will be squeaky clean, and every single port on
the system will be rebuilt with the new binaries, libraries, updated to the
latest versions etc. All that remains would be to rebuild any ports you happen
to prefer to build with custom options, and you're all done.
Works for me, every time.
HTH.
Cheers,
AS
msg41096/pgp0.pgp
Description: PGP signature
4.5-STABLE possible ipnat breakage?
Hi all, I cvsup'd and made my world and kernel via usual means yesterday afternoon, to bring my system to 4.5-STABLE (from a -STABLE version of about two months ago), and I'm experiencing some odd issues. This box is my primary gateway for my home network, and as such, has tircproxy configured to listen on localhost:6999, and has ipnat redirect rules to redirect irc connections through the proxy. Before the upgrade, the transparent proxy worked flawlessly, but as soon as I updated, it started to not work. No ruleset changes. Thinking that maybe some kernel-space structure that tircproxy uses to dig out ipnat information to figure out where to forward the connection changed, I rebuilt tircproxy, and tried again, to no avail. truss'ing the tircproxy process shows it's waiting for accept(), which it comes out of if I do telnet localhost 6999 on the machine, however, when I try to connect from inside my lan to outside it on port 6667 (I have a redirect rule to catch this), it does not come out of the accept(). It looks as tho my connection is somehow being lost in the middle there. I checked, and redirects from outside-in (ie, services on private hosts accessible from outside) work fine, removing the redirect and trying to connect outbound from the lan works (although, I don't have the proxying features of tircproxy when I do that) so I know the rule works, but my connection never gets to tircproxy. Anyone have any ideas as to where to look? Would this have anything to do with the big network stack changes I keep hearing about? Thanks in advance for any responses I may get. -gnp To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
