Re: Need instructions: build kernel on one machine; install onanother
On Thu, 2002-08-15 at 10:59, Ruslan Ermilov wrote: On Thu, Aug 15, 2002 at 09:34:07AM -0500, Stephen Hilton wrote: On 15 Aug 2002 09:22:05 -0500 Larry Rosenman [EMAIL PROTECTED] wrote: On Thu, 2002-08-15 at 09:16, Robin P. Blanchard wrote: from an earlier discussion (initiated by me): You *MUST* use a HOST SYSTEM compiled for the LOWEST processor you want to installworld to. There are a few places (strip for one) that link in /usr/lib/libc.a and NOT the one from /usr/obj/... Thus you can't use an i686 optimized world host box to install a lower optimized /usr/obj to. see also PR i386/30276 from me. Larry, I agree, with FreeBSD having such an edge in performance on older hardware (sorry Intel and AMD :-) this is something to pursue. --snip-- Fix Update build system to NOT include host libc.a in any tools that will be executed on the target system. --snip-- To be honest, I have no brilliant idea how to fix this. One possible solution would be to split build- and install- tools, and always build install-tools on an installing machine, as part of installworld (this should probably be made a special case). The most problematic thing here is strip(1) which is part of binutils; that would mean we would need to compile binutils twice. I am not saying it's impossible, I am saying it is hard. Many things that are used during build are also used during install, like for example, GNU texinfo suite: makeinfo(1) is used during build, install-info(1) -- during install. That also means we need to compile texinfo twice. could we pre-strip those binaries that are installed somehow? Could we modify the binutils build to use the built /usr/obj/usr/lib/libc as a 2nd pass after they are built? Or we could go a lazy way and document that we only support this sort of things if the installed world on a building machine is suitable for CPU of the installing machine. All of my machines here are P6-type, so I usually build world and all kernels on one fast machine and do NFS installs on others. This pretty much is the current state. :-( -- Larry Rosenman http://www.lerctr.org/~ler Phone: +1 972-414-9812 E-Mail: [EMAIL PROTECTED] US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: cvsup Problem?
randy I am getting the following error when I was trying to update my source randy tree: randy Connected to cvsup3.FreeBSD.org randy Server warning: Cannot open randy /usr/local/etc/cvsup/prefixes/FreeBSD.cvs/src/contrib/sendmail/BuildTools/OS/A-UX,v: randy Is anyone else getting this message, and is it being fixed? I've been told it's only on cvsup3. Try changing servers until it is resolved. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: cdrom: device not configured error
Doug, Have you deleted your acd* devices in /dev and used /dev/MAKEDEV to re-create them? The new ATA drivers require this as the minor mode must change. If you did a mergemaster after any upgrade, this should have been taken care of. ls -l /dev/acd* crw-r- 4 root operator 117, 0 Aug 9 11:17 /dev/acd0a crw-r- 4 root operator 117, 0 Aug 9 11:17 /dev/acd0c crw-r- 4 root operator 117, 8 Jul 25 15:46 /dev/acd1a crw-r- 4 root operator 117, 8 Jul 25 15:46 /dev/acd1c Note the major (117) and minor (0) mode of acd0?. If yours does not match, that is likely your problem. R. Kevin Oberman, Network Engineer Energy Sciences Network (ESnet) Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab) E-mail: [EMAIL PROTECTED] Phone: +1 510 486-8634 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: cdrom: device not configured error
--- Kevin Oberman [EMAIL PROTECTED] wrote: Have you deleted your acd* devices in /dev and used /dev/MAKEDEV to re-create them? The new ATA drivers require this as the minor mode must change. If you did a mergemaster after any upgrade, this should have been taken care of. ls -l /dev/acd* crw-r- 4 root operator 117, 0 Aug 9 11:17 /dev/acd0a crw-r- 4 root operator 117, 0 Aug 9 11:17 /dev/acd0c crw-r- 4 root operator 117, 8 Jul 25 15:46 /dev/acd1a crw-r- 4 root operator 117, 8 Jul 25 15:46 /dev/acd1c Note the major (117) and minor (0) mode of acd0?. If yours does not match, that is likely your problem. Yes, this has been done and that is how they look. Any other ideas? Thanks, Doug __ Do You Yahoo!? HotJobs - Search Thousands of New Jobs http://www.hotjobs.com To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
[PATCH] (was: Laptop doesn't free memory so old dmesg can be seen)
Hi, After informations given by Brook Davis (thanks to him :p) I investigate in /sys/kern/subr_prf.c . I just add a bzero() but it fixes the *problem*. I think it doesn't break anything but every comments are welcome. Can the attached patch be commited ? -- Aurélien --- subr_prf.c.old Thu Aug 15 18:59:18 2002 +++ subr_prf.c Thu Aug 15 18:58:45 2002 -855,6 +855,7 msgbufp-msg_size = (char *)msgbufp - cp; } msgbufp-msg_ptr = cp; + bzero(msgbufp-msg_ptr, ((char *)msgbufp - cp)); if (msgbufmapped oldp != msgbufp) msgbufcopy(oldp); msgbufmapped = 1; msg48710/pgp0.pgp Description: PGP signature
Re: [PATCH] (was: Laptop doesn't free memory so old dmesg can be seen)
On Thu, Aug 15, 2002 at 07:03:12PM +0200, Aurélien Nephtali wrote: Hi, After informations given by Brook Davis (thanks to him :p) I investigate in /sys/kern/subr_prf.c . I just add a bzero() but it fixes the *problem*. I think it doesn't break anything but every comments are welcome. Can the attached patch be commited ? You missed the point! This is a feature that some people want. You fix isn't one. IMO, the correct solution would be to mark boots and fix dmesg to only display messages since the last boot unless you specify a flag. -- Brooks -- Any statement of the form X is the one, true Y is FALSE. PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4 msg48711/pgp0.pgp Description: PGP signature
Re: cdrom: device not configured error
Doug wrote: --- Kevin Oberman [EMAIL PROTECTED] wrote: Have you deleted your acd* devices in /dev and used /dev/MAKEDEV to re-create them? The new ATA drivers require this as the minor mode must change. If you did a mergemaster after any upgrade, this should have been taken care of. ls -l /dev/acd* crw-r- 4 root operator 117, 0 Aug 9 11:17 /dev/acd0a crw-r- 4 root operator 117, 0 Aug 9 11:17 /dev/acd0c crw-r- 4 root operator 117, 8 Jul 25 15:46 /dev/acd1a crw-r- 4 root operator 117, 8 Jul 25 15:46 /dev/acd1c Note the major (117) and minor (0) mode of acd0?. If yours does not match, that is likely your problem. Yes, this has been done and that is how they look. Any other ideas? Do you have a slave device and no master on that controller? FreeBSD may not recognize it. Kent Thanks, Doug __ Do You Yahoo!? HotJobs - Search Thousands of New Jobs http://www.hotjobs.com To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message . -- Kent Stewart Richland, WA http://users.owt.com/kstewart/index.html To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: Need instructions: build kernel on one machine; install on another
On Thu, Aug 15, 2002 at 11:16:13AM -0500, Larry Rosenman wrote: On Thu, 2002-08-15 at 10:59, Ruslan Ermilov wrote: On Thu, Aug 15, 2002 at 09:34:07AM -0500, Stephen Hilton wrote: On 15 Aug 2002 09:22:05 -0500 Larry Rosenman [EMAIL PROTECTED] wrote: On Thu, 2002-08-15 at 09:16, Robin P. Blanchard wrote: from an earlier discussion (initiated by me): You *MUST* use a HOST SYSTEM compiled for the LOWEST processor you want to installworld to. There are a few places (strip for one) that link in /usr/lib/libc.a and NOT the one from /usr/obj/... Thus you can't use an i686 optimized world host box to install a lower optimized /usr/obj to. see also PR i386/30276 from me. Larry, I agree, with FreeBSD having such an edge in performance on older hardware (sorry Intel and AMD :-) this is something to pursue. --snip-- Fix Update build system to NOT include host libc.a in any tools that will be executed on the target system. --snip-- To be honest, I have no brilliant idea how to fix this. One possible solution would be to split build- and install- tools, and always build install-tools on an installing machine, as part of installworld (this should probably be made a special case). The most problematic thing here is strip(1) which is part of binutils; that would mean we would need to compile binutils twice. I am not saying it's impossible, I am saying it is hard. Many things that are used during build are also used during install, like for example, GNU texinfo suite: makeinfo(1) is used during build, install-info(1) -- during install. That also means we need to compile texinfo twice. could we pre-strip those binaries that are installed somehow? Could we modify the binutils build to use the built /usr/obj/usr/lib/libc as a 2nd pass after they are built? Or we could go a lazy way and document that we only support this sort of things if the installed world on a building machine is suitable for CPU of the installing machine. All of my machines here are P6-type, so I usually build world and all kernels on one fast machine and do NFS installs on others. This pretty much is the current state. :-( Please see my follow-up to PR i386/30276. Cheers, -- Ruslan Ermilov Sysadmin and DBA, [EMAIL PROTECTED] Sunbay Software AG, [EMAIL PROTECTED] FreeBSD committer, +380.652.512.251Simferopol, Ukraine http://www.FreeBSD.org The Power To Serve http://www.oracle.com Enabling The Information Age msg48713/pgp0.pgp Description: PGP signature
Re: Need instructions: build kernel on one machine; install onanother
[snip] Please see my follow-up to PR i386/30276. I did. Thanks! -- Larry Rosenman http://www.lerctr.org/~ler Phone: +1 972-414-9812 E-Mail: [EMAIL PROTECTED] US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: cvsup Problem?
In article mit.lcs.mail.freebsd-stable/[EMAIL PROTECTED] you write: /usr/local/etc/cvsup/prefixes/FreeBSD.cvs/src/contrib/sendmail/BuildTools/OS/A-UX,v: randy Is anyone else getting this message, and is it being fixed? I've been told it's only on cvsup3. Try changing servers until it is resolved. And did anybody bother to actually *tell* the admin of cvsup3? (In case it wasn't clear, that was a rhetorical question.) -GAWollman -- Garrett A. Wollman | [G]enes make enzymes, and enzymes control the rates of [EMAIL PROTECTED] | chemical processes. Genes do not make ``novelty- Opinions not those of| seeking'' or any other complex and overt behavior. MIT, LCS, CRS, or NSA| - Stephen Jay Gould (1941-2002) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: cvsup Problem?
wollman And did anybody bother to actually *tell* the admin of cvsup3? wollman (In case it wasn't clear, that was a rhetorical question.) I mailed to jdp as I personally don't know who the cvsup3 admin is. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: Loosing network connection repeatedly without an identifiablecause
On Thu, 15 Aug 2002, Mike Tancsa wrote: At 08:17 AM 15/08/2002 -0700, Benjamin Krueger wrote: Check the number of available mbufs. That should show up in dmesg (or logged in kern.* via syslog) as well e.g. /kernel: All mbuf clusters exhausted, please see tuning(7). Yes, I have seen this on a busy system with a badly tuned kernel. Did you know there was trouble with the fxp driver in 4-STABLE a few days ago? I was having similar bizarre problems (ping works in one direction, not the other, etc.) with an fxp interface and it turned out to be a bad cvs commit. It was discussed in this list. See for instance: http://docs.freebsd.org/cgi/getmsg.cgi?fetch=67068+0+current/freebsd-stable M.D. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: cdrom: device not configured error
If I do a boot -verbose at the ok prompt when booting, the cdrom is detected and works fine! (And I do not get the ATA identify retries exceeded error). If I do not boot with -verbose, the cdrom does not work. What gives? Thanks for any and all ideas. Doug __ Do You Yahoo!? HotJobs - Search Thousands of New Jobs http://www.hotjobs.com To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message
Re: Loosing network connection repeatedly without an identifiable cause
At 10:42 AM 15/08/2002 -0700, Michael A. Dickerson wrote: Did you know there was trouble with the fxp driver in 4-STABLE a few days ago? I was having similar bizarre problems (ping works in one Yes, luigi fixed it with a couple of commits. Also, this problem can be worked around by setting kern.ipc.mcl_pool_max=0 ---Mike To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-stable in the body of the message