Re: Problems with fxp0 on T30 with 5.1-RELEASE
On Thu, Jul 17, 2003 at 09:58:00AM -0600, M. Warner Losh wrote: kevin oberman suggested the use of the ps2 utility for thinkpads. i did not have the chance to try it out myself. I also didn't have the time to look at it the last day and a half, but I have time this weekend so I'll definitely investigate. The part that I'm confused about is what's the 'ps2' utility? There's no mention of it in /usr/ports and the only things I can find with google are a DOS program by that name. Is that what you're referring to? If so, where do I get it? the only odd thing is that 5.0 did not have this bug, while 5.1 has it. if it needs fixing is up to you, i am now happy with the current solution. but then, the default bios settings are so stupid that it may not be feasable to do anything about it. This is the part that confuses me as well, but from 5.0 to 5.1 the fxp driver was completely re-written to use a different API so tracking down the specific problem is going to suck :( I also have XP on the laptop (for work stuff, honest! :) and it's working fine as well, so it's definitely a bug in 5.1. I'll also try getting -CURRENT /usr/src off another system and recompiling from scratch to see if that fixes the issue (not having net sucks! :) Dana Lacoste Ottawa, Canada PS: Someone had said that although they might be able to find the bug they'd need the affected hardware. Well, if you're between Toronto and Quebec City I could arrange a field trip for the weekend, but I'm not shipping my laptop to Europe :) ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Problems with fxp0 on T30 with 5.1-RELEASE
ps2 is the IBM utility to adjust BIOS parameters. While these can also be adjusted under Windows, many are never written to CMOS and are lost on re-boot. (Windows re-loads them from the registry.) ps2 will change the values in CMOS memory. I couldn't get ps2 to tell me anything interesting, but I entered the BIOS and put 5,7,9,10,11 for the first 5 PCI IRQ ids and left the last three on auto and ta-da! everything seems to be working. I'm using the GENERIC 5.1-RELEASE kernel so it's using cbb already (someone had asked that) and I can't see that anything else is failing (I don't have wi-fi and I had already disabled serial, IR, and parallel, so USB is the only thing left :) Woo-Hoo! This is the first time I've had 5.x working completely on my laptop. I'm very very appreciative. Thanks everyone! (Now we should probably find out WHY we have to set the IRQ assignments manually, but seeing as it works I'm not in a rush to find out. Feel free to suggest code changes and I'll be a guinea pig though :) Dana Lacoste Ottawa, Canada ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to [EMAIL PROTECTED]
Problems with fxp0 on T30 with 5.1-RELEASE
http://www.freebsdforums.org/forums/showthread.php?s=threadid=10676 This has more info confirming that it's not just me :) Anyways, I'm willing to try anything: I'm installing 5.1 from CD (BSD Mall! YAY!) and am VERY VERY VERY happy that acpi (appears to be) is working perfectly! (I had to retrograde to 4.x because it was so broken with 5.0 and this is my first return to 5.x land since then :) The problem : fxp0: device timeout appears continuously (about every 15 seconds) while interface is UP details : - IBM T30 laptop, Windows says it's an Intel Pro/100 VE - 5.1-RELEASE (can't cvsup, network doesn't work :) - only happens when device is UP (ifconfig down causes the error to go away) - mii appears to work correctly : ifconfig shows the link status correctly instantly (10/100/none, full/half duplex) and it updates when i remove the cable. Also, the mac address shows up correctly. I don't have anything added to the laptop, no cardbus devices, no usb devices, so i'm not sure what's causing the problem, but like I said, I'm willing to try anything out if you have any suggestions :) Dana Lacoste Ottawa, Canada ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to [EMAIL PROTECTED]
Fun and excitement with CURRENT
OK, I've got just 1.5 bugs left, then I'll be all CURRENT-ed :) (I'm using a brand new laptop. Much different than getting BSD working on my desktops and servers :) Things that work : - Making the third mouse button work on an IBM T30 (trackpoint touchpad model) you have to disable the touchpad in the BIOS. - Making CURRENT compile on 4.7 involves NOT using cvsupit to do your cvsup-ing. or using it and editing it to use src-all so you actually get the whole thing. - When compiling a kernel, don't be a smartass and assume you can make your 4.7 KERNCONF file work in 5.0. start over with a 5.0 GENERIC and modify it. You'll feel MUCH less stupid later. - X on a Radeon mobility 7500 is PERFECT. VERY VERY HAPPY! :) - don't use apm and acpi at the same time. UNFORTUNATELY, that last one has a solution : - DON'T USE ACPI. It really causes a problem. If I turn acpi on (in kernel or as a module same result) it will die a horrible, horrible death. I'll get more info (as in, the text it says when it crashes or the text it says when it boots or whatever) after a couple of reboots, but i was so happy it worked without acpi i'm typing this now. debug later :) - Sound is really really weird. pcm doesn't work with the i810 onboard (it works, but sound is all static-y to the point of inaudibility, and the OSS drivers seem to have a problem. I think the oss drivers don't seem to work without ACPI. or at least not so far. It worked once, just not since :) so there ya go, 1.5 bugs. The ACPI took me quite a while to work around because every time it booted with it it crashed the system before I could do anything productive. REALLY ANNOYING :) (Is there a way to tell the kernel not to AUTOLOAD ACPI on boot? That would be SUPER :) Anyways, I'll keep the machine going and hopefully come up with some ACPI patches in the next few days that cause my laptop not to die a horrible death :) (If you have any such patches let me know! :) Thanks for the 5.0! Dana Lacoste Ottawa, Canada To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: 5.0-RELEASE failing on Thinkpad T30
You are missing the src-sys-crypto collection in your supfile. Either use the src-all collection (*strongly* recommended) or else read the examples in /usr/share/examples/cvsup carefully and make sure you are getting all the collections you need. I guess submitting a patch to ports/net/cvsupit that would include this selection would be a better way, eh? :) (i use cvsupit to make the cvsup file, so that's why i didn't even know i was missing something :) Thanks! Dana Lacoste Ottawa, Canada To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
5.0-RELEASE failing on Thinkpad T30
Well, I have a brand new IBM Thinkpad and it shipped with XP so I decided to put FreeBSD on it. 4.7 (after I gave up on 5.0) installs and works FINE. no problems whatsoever. 5.0, when I install it, does a kernel panic, randomly about 5 minutes into the processing of packages. I don't know what it is or why it happens, but it really sucks. I want my 5.0! I tried installing 4.7, cvsup-ing to RELENG_5 (then to '.' cuz there is no RELENG_5. Why not?) but I can't get '.' to compile because gbde keeps failing, looking for rijlaen-alg-fast.c (i think i spelled that right) which isn't there. IN FACT, cvsup-ing didn't put anything in src/sys/crypto at all. So basically I'm staying with 4.7 because I can't find any way at all to get 5.0 to work. Any suggestions for either problem? I think that compiling my own kernel would prevent the panics (between Intel IDE, ACPI, and the P4, a kernel panic from a GENERIC kernel could be understandable, if disappointing) but I can't compile my own kernel because I can't make buildworld (and having a 5.0 kernel with a 4.7 world seems dumb to me :) Dana Lacoste Ottawa, Canada To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
new feature for /etc/security
Greetings all,
In my local source tree, I have a small modification to /etc/security
which I thought would be good to get in the base tree. The attached .diff
allows /etc/security to keep a record of all non-device related files located
in /dev. Many blackhat utilities, and practices include using the /dev
directory as a location to create sniffer logs, suid binaries, and other evil.
By keeping a database similar to /var/log/setuid.today, administrators can be
notified of any changes to /dev. The diff is against -current, however the
functionality is unchanged between -stable and -current.
Is a PR the best way to submit this type of change? I didn't think a
PR was appropriate because I am not submitting a problem, so much as a small
enhancement.
Anybody see any huge holes in my logic?
Thanks,
visigoth
--
Damieon Stark, CCSE
Unix/Network Security Engineer
plug
currently seeking employment
/plug
__
Damieon Stark | Microsoft: Where do you want to go today?
e: [EMAIL PROTECTED] | Linux: Where do you want to go tommorow?
p: 612.382.6945 | FreeBSD/Sun: Are you guys coming or what?
pgp: 0xBE5D0C57 | http://www.sun.com/solaris - The . in .com
pgp.mit.edu | http://www.freebsd.org - The power to serve!
--
--- security.oldSun Sep 2 19:07:07 2001
+++ securitySun Sep 2 19:59:29 2001
@@ -94,6 +94,27 @@
mv ${TMP} ${LOG}/setuid.today || rc=3
fi
+# Show any files in /dev which are not charactor, block
+# device entries, or symlinks.
+
+find /dev -type f -or -type s -or -type p | xargs ls -al ${TMP}
+
+if [ ! -f ${LOG}/devfiles.today ]; then
+ separator
+ echo No ${LOG}/devfiles.today
+ cp ${TMP} ${LOG}/devfiles.today || rc=3
+fi
+
+if ! cmp ${LOG}/devfiles.today ${TMP} /dev/null; then
+ [ $rc -lt 1 ] rc=1
+ separator
+ echo Checking for changes to non-device files in dev:
+ diff -w ${LOG}/devfiles.today ${TMP}
+ mv ${LOG}/devfiles.today ${LOG}/devfiles.yesterday || rc=3
+ mv ${TMP} ${LOG}/devfiles.today || rc=3
+fi
+
+
# Show changes in the way filesystems are mounted
#
[ -n $ignore ] cmd=egrep -v ${ignore#|} || cmd=cat
PGP signature
new /etc/security mod
Greetings all,
I have had a small modification to /etc/security floating around in my
tree for a little while, and thought it would be best to submit it. The
modifications allow the /etc/security script to keep daily track of changes to
all non char/block special files in /dev. Many nefarious activities tend to
occur in /dev due to the often cryptic file names, and the level of
understanding of the average administrator. I have seen boxen which have
large files in /dev due to sniffer/keylogger activities. A daily database is
created in /var/log similar to /var/log/setuid.today and
/var/log/setuid.yesterday. If I am way off my rocker, or somebody sees an
obvious way to improve the method, please let me know.
I realized this should be submitted when I got a report of the mod being used
to detect a rooted box...
The .diff is against -current, but should apply to -stable as well...
Thanks, and keep up all the wonderful work!
visigoth
--
Damieon Stark, CCSE
Unix/Network Security Engineer
plug
currently seeking employment
/plug
__
Damieon Stark | Microsoft: Where do you want to go today?
e: [EMAIL PROTECTED] | Linux: Where do you want to go tommorow?
p: 612.382.6945 | FreeBSD/Sun: Are you guys coming or what?
pgp: 0xBE5D0C57 | http://www.sun.com/solaris - The . in .com
pgp.mit.edu | http://www.freebsd.org - The power to serve!
--
I'll see your DMCA and raise you a First Amendment.
http://www.anti-dmca.org
--
--- security.oldSun Sep 2 19:07:07 2001
+++ securitySun Sep 2 19:59:29 2001
@@ -94,6 +94,27 @@
mv ${TMP} ${LOG}/setuid.today || rc=3
fi
+# Show any files in /dev which are not charactor, block
+# device entries, or symlinks.
+
+find /dev -type f -or -type s -or -type p | xargs ls -al ${TMP}
+
+if [ ! -f ${LOG}/devfiles.today ]; then
+ separator
+ echo No ${LOG}/devfiles.today
+ cp ${TMP} ${LOG}/devfiles.today || rc=3
+fi
+
+if ! cmp ${LOG}/devfiles.today ${TMP} /dev/null; then
+ [ $rc -lt 1 ] rc=1
+ separator
+ echo Checking for changes to non-device files in dev:
+ diff -w ${LOG}/devfiles.today ${TMP}
+ mv ${LOG}/devfiles.today ${LOG}/devfiles.yesterday || rc=3
+ mv ${TMP} ${LOG}/devfiles.today || rc=3
+fi
+
+
# Show changes in the way filesystems are mounted
#
[ -n $ignore ] cmd=egrep -v ${ignore#|} || cmd=cat
PGP signature
Re: new /etc/security mod
Man, I am having a _dumb_ day Didn't realize I had already sent off the original copy... Please disregard... -- Damieon Stark, CCSE Unix/Network Security Engineer plug currently seeking employment /plug __ Damieon Stark | Microsoft: Where do you want to go today? e: [EMAIL PROTECTED] | Linux: Where do you want to go tommorow? p: 612.382.6945 | FreeBSD/Sun: Are you guys coming or what? pgp: 0xBE5D0C57 | http://www.sun.com/solaris - The . in .com pgp.mit.edu | http://www.freebsd.org - The power to serve! -- I'll see your DMCA and raise you a First Amendment. http://www.anti-dmca.org -- PGP signature
Re: unified pwutil library - thoughts?
Hey all... So, by the fact that I got _no_ replies I must surmise that either my message didn't get posted (confirmed posting with archive), or people didn't read it, or _nobody_ thinks it is a good idea ;) but if nobody thinks it is a good idea, I _think_ I would have gotten some flames/feedback. Here goes again... Below is the message I sent about a unified libpwutil, and possibly integrating it into FreeBSD. _ANY_ feedback, positive or negitive would be appreciated as it will help to validate my existance ;) On Sun, Jun 24, 2001 at 09:05:41PM -0500, Damieon Stark wrote: Hey all, Ok, so my thinking goes a little like this: having chpass, vipw, passwd, rpc.yppasswdd et. al. using the same functions (pw_copy, pw_temp, pw_init...) and then, in order to update the database calling a function which does nothing more than execl(pwd_mkdb) and exit seems kinda silly. For a couple of my projects, having a BSD approved way of modifying the master.passwd/spwd.db would have been _very_ nice, so I have created libpwutil. This new lib could be used to (staticly) link all the afore mentioned command line utils, and create a more central way to manage password changes, as well as creating a reasonably nice C interface to changing passwords. If I am way off my rocker, please let me know. I've also created a _new_ pw_mkdb function which does NOT use execl or system. If this sounds like something that people are interested in, I would be happy to work up some doccumentation, and submit it for inclusion/port-dom as well as patches to chpass and crew if people think it is a good idea. Just _looking_ at the Makefiles for things like chpass brought on the urge to make a change... ;) Also, future support for things like NIS+/LDAP could (possibly) be more easily developed. visigoth P.S. the lib will be BSD licenced, NOT GPL'd ;) Visigoth -- Damieon Stark, CCSE Unix/Network Security Engineer plug currently seeking employment /plug __ Damieon Stark | Microsoft: Where do you want to go today? e: [EMAIL PROTECTED] | Linux: Where do you want to go tommorow? p: 612.382.6945 | FreeBSD: Are you guys comming or what? pgp: 0xBE5D0C57 | pgp.mit.edu | http://www.freebsd.org - The power to serve! -- PGP signature
unified pwutil library - thoughts?
Hey all, Ok, so my thinking goes a little like this: having chpass, vipw, passwd, rpc.yppasswdd et. al. using the same functions (pw_copy, pw_temp, pw_init...) and then, in order to update the database calling a function which does nothing more than execl(pwd_mkdb) and exit seems kinda silly. For a couple of my projects, having a BSD approved way of modifying the master.passwd/spwd.db would have been _very_ nice, so I have created libpwutil. This new lib could be used to (staticly) link all the afore mentioned command line utils, and create a more central way to manage password changes, as well as creating a reasonably nice C interface to changing passwords. If I am way off my rocker, please let me know. I've also created a _new_ pw_mkdb function which does NOT use execl or system. If this sounds like something that people are interested in, I would be happy to work up some doccumentation, and submit it for inclusion/port-dom as well as patches to chpass and crew if people think it is a good idea. Just _looking_ at the Makefiles for things like chpass brought on the urge to make a change... ;) Also, future support for things like NIS+/LDAP could (possibly) be more easily developed. visigoth P.S. the lib will be BSD licenced, NOT GPL'd ;) -- Damieon Stark, CCSE Unix/Network Security Engineer plug currently seeking employment /plug __ Damieon Stark | Microsoft: Where do you want to go today? e: [EMAIL PROTECTED] | Linux: Where do you want to go tommorow? p: 612.382.6945 | FreeBSD: Are you guys comming or what? pgp: 0xBE5D0C57 | http://www.freebsd.org - The power to serve! -- PGP signature
VESA crash + softupdt inconsistency
World is 18 Feb 2001 Kernel is 23 Feb 2001 vidcontrol -g 100x37 VESA_800x600 (while building another kernel) made machine hang, beeping continuously, tcp/ip stack was down, ctrl-alt-del unusable. Plugged power off (forgot to ask for ddb. still a freebsd newbie...) At boot, had an 'UNEXPECTED SOFT UPDATE INCONSISTENCY' on /usr (which had softupdates turned on + noatime). manual fsck corrected about 200 summary informations. Attachment: log with the error, the fsck output and dmesg output (Apologies in advance, if a 14Kb attachment is inapropriate for the list) Cheers, --fred System is from 18 Feb 2001 Kernel is from 23 Feb 2001 Boot log: ad0: 39082MB Maxtor 54098H8 [79406/16/63] at ata0-master UDMA66 acd0: DVD-ROM ASUS DVD-ROM E608 at ata0-slave using UDMA33 Waiting 10 seconds for SCSI devices to settle Mounting root from ufs:/dev/ad0s4a WARNING: / was not properly dismounted swapon: adding /dev/ad0s4b as swap device Automatic boot in progress... /dev/ad0s4a: 2402 files, 45564 used, 4019 free (267 frags, 469 blocks, 0.5% frag mentation) cd0 at ahc0 bus 0 target 5 lun 0 cd0: YAMAHA CRW8424S 1.0f Removable CD-ROM SCSI-2 device cd0: 5.000MB/s transfers (5.000MHz, offset 15) cd0: cd present [1 x 2048 byte records] /dev/ad0s4f: INCORRECT BLOCK COUNT I=2595990 (2256 should be 208) (CORRECTED) /dev/ad0s4f: UNALLOCATED I=2937264 OWNER=root MODE=100644 /dev/ad0s4f: SIZE=7872 MTIME=Feb 25 17:49 2001 /dev/ad0s4f: FILE=/obj/usr/src/sys/AGLAE/modules/usr/src/sys/modules/vpo/immio.o /dev/ad0s4f: UNEXPECTED SOFT UPDATE INCONSISTENCY; RUN fsck MANUALLY. /dev/ad0s4e: 966 files, 6601 used, 13214 free (230 frags, 1623 blocks, 1.2% frag mentation) THE FOLLOWING FILE SYSTEM HAD AN UNEXPECTED INCONSISTENCY: ufs: /dev/ad0s4f (/usr) Automatic file system check failed . . . help! Enter full pathname of shell or RETURN for /bin/sh: # fsck /dev/ad0s4f ** /dev/ad0s4f ** Last Mounted on /usr ** Phase 1 - Check Blocks and Sizes ** Phase 2 - Check Pathnames ** Phase 3 - Check Connectivity ** Phase 4 - Check Reference Counts ** Phase 5 - Check Cyl groups FREE BLK COUNT(S) WRONG IN SUPERBLK SALVAGE? [yn] y SUMMARY INFORMATION BAD SALVAGE? [yn] y ALLOCATED FRAG 926933 MARKED FREE ALLOCATED FRAG 926934 MARKED FREE ALLOCATED FRAG 927080 MARKED FREE ALLOCATED FRAG 927081 MARKED FREE ALLOCATED FRAG 927082 MARKED FREE ALLOCATED FRAG 927083 MARKED FREE ALLOCATED FRAG 927084 MARKED FREE ALLOCATED FRAG 927085 MARKED FREE ALLOCATED FRAG 927086 MARKED FREE ALLOCATED FRAG 927087 MARKED FREE ALLOCATED FRAG 927104 MARKED FREE ALLOCATED FRAG 927105 MARKED FREE ALLOCATED FRAG 927106 MARKED FREE ALLOCATED FRAG 927107 MARKED FREE ALLOCATED FRAG 927108 MARKED FREE ALLOCATED FRAG 927109 MARKED FREE ALLOCATED FRAG 927110 MARKED FREE ALLOCATED FRAG 927112 MARKED FREE ALLOCATED FRAG 927113 MARKED FREE ALLOCATED FRAG 927114 MARKED FREE ALLOCATED FRAG 927115 MARKED FREE ALLOCATED FRAG 927116 MARKED FREE ALLOCATED FRAG 927117 MARKED FREE ALLOCATED FRAG 927118 MARKED FREE ALLOCATED FRAG 927119 MARKED FREE ALLOCATED FRAG 927120 MARKED FREE ALLOCATED FRAG 927121 MARKED FREE ALLOCATED FRAG 927122 MARKED FREE ALLOCATED FRAG 927123 MARKED FREE ALLOCATED FRAG 927124 MARKED FREE ALLOCATED FRAG 927125 MARKED FREE ALLOCATED FRAG 927126 MARKED FREE ALLOCATED FRAG 927127 MARKED FREE BLK(S) MISSING IN BIT MAPS SALVAGE? [yn] y ALLOCATED INODE 238340 MARKED FREE ALLOCATED FRAG 4226888 MARKED FREE ALLOCATED FRAG 4226889 MARKED FREE ALLOCATED FRAG 4226890 MARKED FREE ALLOCATED FRAG 4226891 MARKED FREE ALLOCATED FRAG 4226892 MARKED FREE ALLOCATED FRAG 4226893 MARKED FREE ALLOCATED FRAG 4226894 MARKED FREE ALLOCATED INODE 2595976 MARKED FREE ALLOCATED INODE 2595977 MARKED FREE ALLOCATED INODE 2595978 MARKED FREE ALLOCATED INODE 2595979 MARKED FREE ALLOCATED INODE 2595980 MARKED FREE ALLOCATED INODE 2595981 MARKED FREE ALLOCATED INODE 2595982 MARKED FREE ALLOCATED INODE 2595983 MARKED FREE ALLOCATED INODE 2595984 MARKED FREE ALLOCATED INODE 2595985 MARKED FREE ALLOCATED INODE 2595986 MARKED FREE ALLOCATED INODE 2595987 MARKED FREE ALLOCATED INODE 2595988 MARKED FREE ALLOCATED INODE 2595989 MARKED FREE ALLOCATED INODE 2595990 MARKED FREE ALLOCATED INODE 2595991 MARKED FREE ALLOCATED INODE 2595992 MARKED FREE ALLOCATED INODE 2595993 MARKED FREE ALLOCATED INODE 2595994 MARKED FREE ALLOCATED INODE 2595995 MARKED FREE ALLOCATED INODE 2595996 MARKED FREE ALLOCATED FRAG 10752925 MARKED FREE ALLOCATED FRAG 10752926 MARKED FREE ALLOCATED FRAG 10945763 MARKED FREE ALLOCATED FRAG 10945764 MARKED FREE ALLOCATED FRAG 10945765 MARKED FREE ALLOCATED FRAG 10945766 MARKED FREE ALLOCATED FRAG 10945767 MARKED FREE ALLOCATED FRAG 11212226 MARKED FREE ALLOCATED FRAG 11212227 MARKED FREE ALLOCATED FRAG 11212228 MARKED FREE ALLOCATED FRAG 11212229 MARKED FREE ALLOCATED FRAG 11212230 MARKED FREE ALLOCATED FRAG 11212672 MARKED FREE ALLOCATED FRAG 11212673 MARKED
Re: Make kernel fail in modules after upgrade 4.2 - 5.0
On Sun, 18 Feb 2001, Pete Carah wrote: This may relate to a commit about noon (PST) today fixing a different problem. I'm just waiting it out :-) Oh. I'll just wait too. Welcome to "current" where (especially lately) about half the time things don't 'make'... (I'm trying to recompile my kernel after recovering from the libc circus, trying to prevent some new panics) Yeah. I don't know what I smoked yesterday to decide to go 'current' (In fact, it is because mozilla-0.8 doesn't compile/run on 4.2, but reportedly works on 5.0. I bet I just replaced a small set of problems by a larger one...). On the bright side, this will keep me entertained for weeks :-) Thanks for the reply, Cheers, --fred To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: kernel modules broken (kmod.mk?)
On Mon, 19 Feb 2001, Pierre Y. Dampure wrote:
"Andrey A. Chernov" wrote:
Recent -current, 'make' fails ('make depend' works), I got this for
_every_ module:
ld -r -o 3dfx.kld tdfx_pci.o
/usr/libexec/elf/ld: cannot open tdfx_pci.o: No such file or directory
*** Error code 1
Stop in /usr/src/sys/modules/3dfx.
I think this might be related to Peter Wemm's last commit on
src/sys/conf/kmod.mk, which seems to remove the .depend file at the start
of the make. If I revert the change, modules compile OK
After I commented out line 161 of 'kmod.mk' ( @rm -f .depend ) and removed
the linux module from src/sys/modules/Makefile the compile went OK.
Cheers,
--fred
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
Re: name resolution problems
On Mon, 19 Feb 2001, Wesley Morgan wrote: Since the big shake-up with -current, I find that mozilla and galeon can no longer function (both up to date), but lynx has no problems. Mozilla seems stuck resolving hostnames, yet tcpdump shows no traffic and truss indicates that it is simply looping around a poll(). The biggest difference between lynx and mozilla in terms of name resolution is that mozilla is linked against libc_r... Could there be some problem here? Is anyone else seeing this? I see that too. I removed ~/.mozilla, and it works well, but only once. At next launch it get stuck in name resolution again, until I remove ~/.mozilla. Weird. Cheers, --fred To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Make kernel fail in modules after upgrade 4.2 - 5.0
Hi, What I did: cvsup from 4.2-STABLE to 5.0-CURRENT make buildworld make installworld mergemaster -- All this went okay Then updated my kernel config file and compiled the kernel. (I still haven't rebooted the machine) The build fails in: === 3dfx ld -r -o 3dfx.kld tdfx_pci.o /usr/libexec/elf/ld: cannot open tdfx_pci.o: No such file or directory *** Error code 1 Stop in /usr/src/sys/modules/3dfx. *** Error code 1 Stop in /usr/src/sys/modules. *** Error code 1 Stop in /usr/src/sys/compile/AGLAE. 3dfx seems to be the first module. If I try to build it by hands: su-2.03# pwd /usr/src/sys/modules/3dfx su-2.03# make clean rm -f setdef0.c setdef1.c setdefs.h setdef0.o setdef1.o 3dfx.ko 3dfx.kld tdfx_pci.o @ machine symb.tmp tmp.o opt_tdfx.h bus_if.h device_if.h pci_if.h su-2.03# make Warning: Object directory not changed from original /usr/src/sys/modules/3dfx @ - /usr/src/sys machine - /usr/src/sys/i386/include echo "#define TDFX_LINUX" opt_tdfx.h perl @/kern/makeobjops.pl -h @/kern/bus_if.m perl @/kern/makeobjops.pl -h @/dev/pci/pci_if.m perl @/kern/makeobjops.pl -h @/kern/device_if.m ld -r -o 3dfx.kld tdfx_pci.o /usr/libexec/elf/ld: cannot open tdfx_pci.o: No such file or directory *** Error code 1 Stop in /usr/src/sys/modules/3dfx. su-2.03# ls @ bus_if.hmachine pci_if.h Makefiledevice_if.h opt_tdfx.h If I try to build any other module, I have the same error. Looks like my makefiles are in bad shape and forget to build the object files. What can I do ? Thanks in advance, --fred To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
That des nolinks thang..
Could somebody repost the flag intended for make.conf to keep the des libs from taking over libcrypt's links? Wrote it down, deleted it, lost the paper, and comenced feeling dumb ;) Thanks for all of the hard work everybody! Damieon Stark To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
