5.2-BETA and related ports issues
Things are actually looking pretty good at this point; I'm probably going to move from 4.9-STABLE to 5.2-RELEASE on my main home server, but I'm seeing the following with 5.2-BETA at this point: I'm running 5.2-BETA cvsup'd at about 9 PM 11/25 on two systems; one is a Supermicro P6DGH, dual PIII/850, SCSI disks, Compaq/Intel Pro100+ NIC; the other is a Toshiba 8100 Tecra using the built-in 3Com 10/100 port in the docking station (xl driver). The P6DGH is running netatalk with the (apparently) as yet uncommitted patch to the netatalk kernel interface, which appears to be running fine. I am not seeing any IP lockups (just finished a buildword/buildkernel/ installkernel/installword cycle with the portable NFS mounting the P6DGH). The following appeared before 5.2-BETA but are continuing with that version. On the portable I'm getting the following lock order reversal: Nov 28 10:32:33 mikes-port kernel: lock order reversal Nov 28 10:32:33 mikes-port kernel: 1st 0xc340eb00 pcm0 (sound softc) @ /usr/src/sys/dev/sound/pci/ds1.c:734 Nov 28 10:32:33 mikes-port kernel: 2nd 0xc340e740 pcm0:play:0 (pcm channel) @ /usr/src/sys/dev/sound/pcm/channel.c:440 Nov 28 10:32:33 mikes-port kernel: Stack backtrace: The related parts of dmesg are as follows (an improvement from 5.1-CURRENT, where the AC97 driver didn't load): pcm0: port 0xbf3c-0xbf3f,0xbf40-0xbf7f mem 0xefdf8000-0x efdf irq 11 at device 12.0 on pci0 ds1: setmap (226000, 3de4), nseg=1, error=0 pcm0: pcm0: Codec features headphone, 18 bit DAC, 18 bit ADC, 5 bit master volume, AKM 3D Audio pcm0: Primary codec extended features variable rate PCM, AMAP pcm0: sndbuf_setmap 135db000, 1000; 0xc3407000 -> 135db000 pcm0: sndbuf_setmap 13579000, 1000; 0xc3405000 -> 13579000 pcm0: sndbuf_setmap 13134000, 1000; 0xc340 -> 13134000 pcm0: sndbuf_setmap 134b2000, 1000; 0xc341e000 -> 134b2000 pcm0: sndbuf_setmap 134d, 1000; 0xc341c000 -> 134d pcm0: sndbuf_setmap 1348e000, 1000; 0xc341a000 -> 1348e000 postgreSQL startup called twice On both systems I'm running postgreSQL7 from ports. In both cases the pgctl (the startup script) is called twice, and obviously it fails the second time. It is called both by /etc/rc.d/localdaemons and by /etc/rc.d/localpkg. I haven't looked any deeper than that, yet. On the portable the IP number, netmask, and router address are set in /etc/rc.conf. Both /etc/rc.d/netoptions and /etc/rc.d/network3 appear to be execuring (I see 'Additional TCP options:" twice) and one of them is trying to reset the router address set by rc.conf, resulting in an error. The plus side is that a lot of the ACPI errors I used to get on the admittedly wierd P6DGH (11 PCI slots, onboard I2O, etc) have gone away. I'm not adding anything else at this point, since I don't know if these have been already reported (or fixed), but I can provide other info if necessary. Mike Squires UN*X at home since 1986 ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: 5.2-BETA + netatalk = crash
> On Thursday 27 November 2003 06:35 pm, Leo Bicknell wrote: > > Since applying your patch I'd have IPv4 stop working 4 times. No panic, > > no console errors, just IPv4 traffic no longer does anything. Can't > > forward through the box. Can't ping the box, can't do anything. > > Logging in on console everything appears fine, and a reboot clears it > > up. > > > > I just reverted to the old kernel, we'll see if it happens again. > > The change I gave you should have nothing to do with IPv4. There is a change > pending commit that appears to fix certain system lockups. I completed a make buildworld on a portable which NFS mounts /usr/src on the 5.2-BETA box which also runs netatalk, no obvious problems. Network hardware is a Compaq/Intel Pro100+ card using the fxp driver. Mike Squires ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: 5.x DOS against NFS server
Guy Van Sanden > > I just ran nmap ... > Nessus has the same effect. When running "nmap (nmap 2.53 on a 4.9-STABLE box)) against a 5.2-BETA host on the host I see Nove 27 13:06:24 mikes sm-mta[483]: NOQUEUE: SYSERR(root): getrequests: accept: Software caused connection abort Nove 27 13:06:24 mikes nfsd[392]: accept failed: Software caused connection abort between messages about response rate limits to nmap queries. On the client running 5.1-CURRENT with "bonnie" using a 100MB file on a volume NFS mounted from the 5.2-BETA server there are no log messages and no obvious error messages; "bonnie" finishes normally. Mike Squires ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: 5.2-BETA + netatalk = crash
Sam Leffler > > > > On Wednesday 26 November 2003 06:51 am, Michael L. Squires wrote: > > > > On my dual CPU P6DGH the 11/22 cvsup of 5.2-BETA and netatalk crashes > > > > on boot. Stopping netatalk from starting stops the crash. > > > > > > > > I wasn't able to catch any crash information, but am currently > > > > recompiling with sources from last night to see if it's repeatable. > > > > > > Please try the attached patch, it should fix at least one problem in the > > > netatalk code (Leo this should fix the panics you've seen during > > > shutdown). Michael, hopefully this will also fix your problem--whatever > > > it is. The patch certainly fixes the panic, thanks. I'm running a make buildworla, etc., on my notebook which NFS mounts the box running netatalk, which should trigger the IPv4 bug if I have it. Mike Squires ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
5.2-BETA + netatalk = crash
On my dual CPU P6DGH the 11/22 cvsup of 5.2-BETA and netatalk crashes on boot. Stopping netatalk from starting stops the crash. I wasn't able to catch any crash information, but am currently recompiling with sources from last night to see if it's repeatable. Mike Squires ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: How to fix this in 5.1-REL??
> /usr/src/contrib/libstdc++/libsupc++/unwind-cxx.h:41:20: unwind.h: No such= > file or directory > mkdep: compile failed > *** Error code 1 > =20 > Stop in /usr/src/gnu/lib/libstdc++. > *** Error code 1 I'm running 5.1-CURRENT now, but I was able to build 5.1-RELEASE-p10. I suspect you need to re-cvsup your sources, which I've had to do several times. Mike Squires ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
gcc ICE building R-letter is gone with gcc 3.3.3
The internal compiler error compiling R-letter (R-1.8.0) under 5-CURRENT went away when gcc was updated to 3.3.3. Mike Squires ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: DRM error messages - easy way to turn them off?
> I've gone ahead and made it a debug message again in -current. It's Worked. I now get kernel: info: [drm] Initialized rdeon 1.9.0 20020828 on minor 0 kernel: drm0: [MPSAFE] Thanks! Mike Squires SM P6DGH/AiW 7200 ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
DRM error messages - easy way to turn them off?
> > On Sat, 25 Oct 2003 04:50:30 +0900, Shin-ichi Yoshimoto wrote: > > > I'm using a radeon DRM module. > > > > Now DRM work fine. But some kernel messages found my log like this. > > > > kernel: error: [drm:pid38523:radeon_ioctl] *ERROR* can't find > > authenticator > > > > Is this OK ? > Is there any simple way to turn this error off? These are appearing frequently enough to fill /var/log/messages several times a day, although I haven't seen any other side effects. Mike Squires SM P6DGH/AiW 7200 ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: FreeBSD-CURRENT kernel hangs on starting KDE
> System: FreeBSD 5.1-CURRENT i386 > > Relevant hardware config: > Mainboard : ASUS P4S533-E (SIS645DX north, SIS962/L south bridges) > Memory : PC2700 512MB DDR RAM (memtest86 tested) > VGA card : ATI RADEON 9000 PRO I had similar lockups with a SuperMicro P6DGH (dual PIII/850)/ATI All-in Wonder (RADEON 7200)/Intel Pro100+ which were solved by moving the Ethernet card to another IRQ (it was configured to share 11 by default, moved it to 9). The system would just stop working - mouse frozen, no keyboard response, etc. I use KDE, also. Mike Squires ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Success compiling R v 1.8.0
The "R" statistics port (ports/math/R-letter) is at version 1.7.0. Being the impatient type I tried compiling the recently released 1.8.0 using the same Makefile (with 1.7 changed to 1.8) and a new MD5 value, and had no new problems compiling. R v.1.7 and v 1.8 trigger a bug in GCC (internal compiler error) and the workaround is to wait for the bug to trigger, then compile that particular piece of code using "-O2" rather than "-O", and then execute "make" again. Mike Squires ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: DRM related error
Janet Sullivan > > Shin-ichi Yoshimoto wrote: > > Now DRM work fine. But some kernel messages found my log like this. > > > > kernel: error: [drm:pid38523:radeon_ioctl] *ERROR* can't find > > authenticator > > DRM is working again for me as well, but I'm also seeing a lot of these > error messages with my Radeon 9000 Mobility. Me to - 752 in the past few hours since rebooting. Mike Squires ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
drm errors with All-in-Wonder Radeon 7200
Since the drm updates to 5-CURRENT I'm getting multiple instances of the following at boot (and not trap 12 :-) ): from dmesg drm0: port 0xc800-0xc8ff mem 0xff48-0xff4f, 0xe80-0xefff irq 11 at device 0.0 on pci1 info: [drm] AGP at 0xf800 64MB info: [drm] Initalized radeon 1.9.0 20020828 on minor 0 drm0: [MPSAFE] error: [drm:pid619:radeon_ioctl] *ERROR* can't find authenticator ... error: [drm:pid720:radeon_ioctl] *ERROR* can't find authenticator which eventually stop. I'm currently composing this message using KDE 3.1.4 using that display. The card is a Radeon 7200 (original All-in-Wonder) which I thought was 32MB, but I could be wrong. Motherboard is an old SM P6DGH dual PIII/850 which has i20 on-board (not in use) and dual PCI card busses with a total of 11 slots. The AGP slot is allegedly AGP 1X/2X. Mike Squires ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: fresh -current trap
> > Fatal trap 12: page fault while in kernel mode > > I just committed a fix for this to CVS. Sorry for the trouble. > > -- > Eric Anholt[EMAIL PROTECTED] This looks like the bug that was causing me trouble. Mike Squires ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
P6DGH and CURRENT as of 2 PM EST 10/24 panics
-CURRENT as of 10/17 is not a problem, kernel just compiled with SMP support today after cvsup at about 2 PM EST (10/24) panics. Nothing shows up in the logs (of course?). I'm recompiling with no SMP support and will report if that dies also. Hardware is the oddball SM P6DGH dual PIII/850 with bridged PCI busses and onboard Adaptec SCSI support, plus i2o support that is not used. The 10/18 kernel boots with a lot of ACPI errors, but so far they have had no effect. Mike Squires ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"
10/23 cvsup buildworld failure
Tinderbox > /vol/vol0/users/des/tinderbox/CURRENT/amd64/amd64/src/lib/libfetch/common.c:60: > error: initializer element is not constant Under 5.1-CURRENT or i386 cvsuped on 10/23 I get the same failure, with the additional error message /usr/src/lib/libfetch/common.c:58: error: 'EAINONAME' undeclared here (not in a function) Other error messages are similar or the same. I have re-cvsup'd and am trying again. Mike Squires ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "[EMAIL PROTECTED]"