https://bugzilla.novell.com/show_bug.cgi?id=336277#c1 --- Comment #1 from Daniel Mader 2007-10-24 05:42:43 MST --- Created an attachment (id=180210) --> (https://bugzilla.novell.com/attachment.cgi?id=180210) stdout of lspci command -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277#c3 --- Comment #3 from Daniel Mader 2007-10-24 05:43:37 MST --- Created an attachment (id=180212) --> (https://bugzilla.novell.com/attachment.cgi?id=180212) stdout of lshal command -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 Daniel Mader changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #180211|stdout of lspci command |stdout of lsmod command description| | Attachment #180211|application/octet-stream |text/plain mime type| | -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277#c6 --- Comment #6 from Daniel Mader 2007-10-25 02:23:23 MST --- Hello we have checked the BIOS features thoroughly and there is no such option. Also, remember that it is an Intel machine, not AMD---but their equivalent XD-Bit is also not present... Maybe it is the P4 HT hyperthreading feature? It is really weird to see all those segfaults during boot time. The x86 version runs fine on other machines so it does not seem likely to me that there is a general bug somewhere. The machine is intended to become a terminal server/simulation machine because of its strong processor... Would be very nice to get some hints on how to go on from here! -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277#c8 Mark Gordon changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mtgordon@novell.com Status|NEW |NEEDINFO Info Provider| |danielstefanmader@web.de --- Comment #8 from Mark Gordon 2007-10-26 12:56:43 MST --- Does the boot get as far as logging? If so, perhaps you could salvage and attach /var/log/messages Do you have any better luck booting with maxcpus=1 ? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277#c10 --- Comment #10 from Daniel Mader 2007-10-30 08:01:24 MST --- Created an attachment (id=181225) --> (https://bugzilla.novell.com/attachment.cgi?id=181225) /var/log/boot.msg after trying to boot into runlevel 3 with safe settings -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277#c12 --- Comment #12 from Mark Gordon 2007-10-30 15:19:04 MST --- Well, installing an X pattern would pull in the enhanced_base pattern, which would pull in microcode_ctl. It's not abundantly clear to me why you suspect that's the problem, but you've seen more boot messages than just those attached. If you think that's the problem, you could mess with the symlinks so as to avoid starting microcode.ctl at whichever runlevel you feel like testing (or just uninstall the package temporarily, or edit the init script so that it's a noop, or...). Can you give that a try? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 Mark Gordon changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|bnc-team-screening@forge.provo.novell.com |bwalle@novell.com Severity|Blocker |Critical -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277#c15 Daniel Mader changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|danielstefanmader@web.de | --- Comment #15 from Daniel Mader 2007-11-05 09:10:37 MST --- The box is a desktop computer which has a firewire interface. So if you could point me to some webpage which describes how to retrieve the oops message I could maybe provide that info, too. For the time being I am going with a remote ssh-session: * Output for step 3: ip6_tables: (C) 2000-2006 Netfilter Core Team ip_tables: (C) 2000-2006 Netfilter Core Team audit(1193834991.074:3): audit_pid=3498 old=0 by auid=4294967295 eth1: no IPv6 routers present IA-32 Microcode Update Driver: v1.14a * Output for step 5: chambertin:~ # /sbin/microcode_ctl -u /sbin/microcode_ctl: writing microcode (length: 320560) /sbin/microcode_ctl: microcode successfuly written to /dev/cpu/microcode * /var/log/messages shows something interesting, though: Nov 5 16:55:55 chambertin firmware.sh[24192]: Cannot find firmware file 'intel-ucode/0f-02-09' Nov 5 16:55:55 chambertin kernel: IA-32 Microcode Update Driver: v1.14a Nov 5 16:55:55 chambertin firmware.sh[24196]: Cannot find firmware file 'intel-ucode/0f-02-09' * The system is still up after that remotely and I can change to another user: Nov 5 16:57:55 chambertin su: (to mader) root on /dev/pts/0 * Yet, when trying to log in locally, I get this (from /var/log/messages): Nov 5 17:01:36 chambertin dbus-daemon: nss_ldap: reconnected to LDAP server ldap://rioja.imtek.uni-freiburg.de after 1 attempt Nov 5 17:01:41 chambertin kernel: BUG: unable to handle kernel NULL pointer dereference at virtual address 00000018 Nov 5 17:01:41 chambertin kernel: printing eip: Nov 5 17:01:41 chambertin kernel: c01784e4 Nov 5 17:01:41 chambertin kernel: *pde = 00000000 Nov 5 17:01:41 chambertin kernel: Oops: 0002 [#1] Nov 5 17:01:41 chambertin kernel: SMP Nov 5 17:01:41 chambertin kernel: last sysfs file: /devices/pci0000:00/0000:00:00.0/irq Nov 5 17:01:41 chambertin kernel: Modules linked in: joydev microcode firmware_class nfs lockd nfs_acl sunrpc autofs4 iptable_filter ip_tables ip6table_filter ip6_tables x_tables ipv6 af_packet snd_pcm_oss snd_mixer_oss snd_seq snd_seq_device apparmor loop dm_mod ohci1394 ieee1394 snd_intel8x0 snd_ac97_codec e100 ac97_bus snd_pcm snd_timer parport_pc mii usb_storage ide_core usbhid snd rtc_cmos ns558 gameport button intel_agp agpgart soundcore snd_page_alloc i2c_i801 i2c_core iTCO_wdt hid rtc_core ff_memless rtc_lib parport iTCO_vendor_support sr_mod shpchp pci_hotplug cdrom sg sd_mod ehci_hcd uhci_hcd usbcore edd ext3 mbcache jbd fan ata_piix libata scsi_mod thermal processor Nov 5 17:01:41 chambertin kernel: CPU: 1 Nov 5 17:01:41 chambertin kernel: EIP: 0060:[<c01784e4>] Tainted: G N VLI Nov 5 17:01:41 chambertin kernel: EFLAGS: 00210282 (2.6.22.9-0.4-default #1) Nov 5 17:01:41 chambertin kernel: EIP is at do_path_lookup+0x1c/0x1cc Nov 5 17:01:41 chambertin kernel: eax: dffb1030 ebx: dfe06cc0 ecx: 08081ec0 edx: d0744000 Nov 5 17:01:41 chambertin kernel: esi: 00000000 edi: b770f0ae ebp: b770f0ae esp: c80f7f80 Nov 5 17:01:41 chambertin kernel: ds: 007b es: 007b fs: 00d8 gs: 0033 ss: 0068 Nov 5 17:01:41 chambertin kernel: Process drkonqi (pid: 24718, ti=c80f6000 task=dffb1030 task.ti=c80f6000) Nov 5 17:01:41 chambertin kernel: Stack: d0744000 c01773ff d0744000 d0744000 08081ec0 b770f0ae c0178e15 00000000 Nov 5 17:01:41 chambertin kernel: b770f0ce 08081ec0 b6c12d06 c80f6000 c016ed73 00000000 08081ec0 00000004 Nov 5 17:01:41 chambertin kernel: b6c1cff4 08081ec0 b6c12d06 bfd0a2e8 ffffffda 0000007b 0000007b c02c0000 Nov 5 17:01:41 chambertin kernel: Call Trace: Nov 5 17:01:41 chambertin kernel: [<c01773ff>] getname+0x59/0xab Nov 5 17:01:41 chambertin kernel: [<c0178e15>] __user_walk_fd+0x2f/0x40 Nov 5 17:01:41 chambertin kernel: [<c016ed73>] sys_fchmodat+0x30/0xbb Nov 5 17:01:41 chambertin kernel: [<c02c0000>] unix_bind+0x176/0x28e Nov 5 17:01:41 chambertin kernel: ======================= Nov 5 17:01:41 chambertin kernel: Code: ba 01 00 00 00 83 c4 14 89 d0 5b 5e 5f 5d c3 55 57 89 c7 56 53 83 ec 08 8b 74 24 1c 89 14 24 64 a1 00 20 3c c0 8b 98 60 04 00 00 <c7> 46 18 01 00 00 00 89 4e 14 c7 46 1c 00 00 00 00 8b 04 24 80 Nov 5 17:01:41 chambertin kernel: EIP: [<c01784e4>] do_path_lookup+0x1c/0x1cc SS:ESP 0068:c80f7f80 Nov 5 17:01:49 chambertin kernel: MCE: The hardware reports a non fatal, correctable incident occurred on CPU 1. Nov 5 17:01:49 chambertin kernel: Bank 1: 8002042800000134 Nov 5 17:01:50 chambertin kdm[3482]: X server for display :0 terminated unexpectedly Nov 5 17:01:51 chambertin powersaved[3687]: WARNING (filter_function:201) Hal service stopped. Battery information no longer available * After this, the system become unusable... -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User bwalle@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c16 Bernhard Walle changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |danielstefanmader@web.de --- Comment #16 from Bernhard Walle 2007-12-13 08:18:59 MST --- Sorry for not responding on that but for a long time. I was on vacation first and then busy with other stuff. Did you try the latest upstream kernel? You can find packages at ftp://ftp.gwdg.de/pub/linux/suse/ftp.suse.com/projects/kernel/kotd/HEAD If not, this needs to be debugged, of course. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User bwalle@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c18 --- Comment #18 from Bernhard Walle 2007-12-14 05:06:59 MST --- The HEAD kernels have very few patches and are very close to upstream. Most patches in the 10.3 kernel are backports from later upstream versions. The days were the SUSE kernel had lots of patches are over. That's only SLES since lots of backports and some additional features like EVMS or LKCD are necessary, but not openSUSE. However, I didn't advice to replace the distribution kernel with the HEAD kernel, just test the microcode updater. You can install the kernel RPM with $ rpm -ivh kernel-default...rpm RPM can handle installation of multiple versions of the same package. You just have to check the default kernel in /boot/grub/menu.lst and after the test you can remove it with $ rpm -e kernel-default-version again. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User danielstefanmader@web.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c20 --- Comment #20 from Daniel Mader 2007-12-18 06:23:53 MST --- Created an attachment (id=188048) --> (https://bugzilla.novell.com/attachment.cgi?id=188048) /var/log/boot.msg after regular boot see lines 544 ff -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 Bernhard Walle changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #188048|application/octet-stream |text/plain mime type| | -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User bwalle@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c23 --- Comment #23 from Bernhard Walle 2007-12-18 07:26:49 MST --- (In reply to comment #22 from Daniel Mader)

We have a live CD floating around and I'll try with it. If that contains the microcode-package I'll happily provide the requested info, otherwise I'm afraid I won't be able to help any further.

Well, just mount the SUSE partition from CD and execute the SUSE stuff there. But don't chroot since then it cannot load the kernel module properly. I'd be only interested in the kernel part, and since the microcode kernel part is mainline, I'm quite sure the Debian/Ubuntu CDs contain it. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User danielstefanmader@web.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c25 Daniel Mader changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|danielstefanmader@web.de | --- Comment #25 from Daniel Mader 2008-01-07 03:20:31 MST --- Hello, both on Knoppix 5.2 (and also on the earlier 5 only) and on Debian 4.0, the microcode module is not loaded by default. Doing so yields this in the tail of dmesg: EXT3-fs: mounted filesystem with ordered data mode. NET: Registered protocol family 17 e100: eth1: e100_watchdog: link up, 100Mbps, full-duplex e100: eth1: e100_watchdog: link up, 100Mbps, full-duplex IA-32 Microcode Update Driver: v1.14a /sbin/microcode_ctl does not exist on the systems at all. Mounting the openSUSE 10.3 system partition and calling it from there brings an error: /sbin/microcode_ctl: /lib/tls/i686/cmov/libc.so.6: version `GLIBC_2.4' not found (required by ./sbin/microcode_ctl) So obviously the Debian systems don't have issues with microcode because they don't use it at all. I can only conclude that there is something wrong with the provided microcode_ctl. Thank you for your efforts on that! -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User danielstefanmader@web.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c27 --- Comment #27 from Daniel Mader 2008-01-07 04:22:59 MST --- OK, I did as you told me. Turns out that Debian 4.0 has pretty much the same problems once microcode and microcode_ctl is active: * Loading the microcode module: eth2: no IPv6 routers present IA-32 Microcode Update Driver: v1.14a microcode: CPU0 updated from revision 0x21 to 0x2e, date = 08112004 microcode: CPU1 updated from revision 0x21 to 0x2e, date = 08112004 IA-32 Microcode Update Driver: v1.14a * the shell suddently gets flooded with a trace, which appears in the last 33 lines of dmesg as follows: eth2: no IPv6 routers present IA-32 Microcode Update Driver: v1.14a microcode: CPU0 updated from revision 0x21 to 0x2e, date = 08112004 microcode: CPU1 updated from revision 0x21 to 0x2e, date = 08112004 IA-32 Microcode Update Driver: v1.14a BUG: unable to handle kernel NULL pointer dereference at virtual address 00000000 printing eip: f8994b12 *pde = 00000000 Oops: 0002 [#1] SMP Modules linked in: microcode ppdev lp button ac battery ipv6 dm_snapshot dm_mirror dm_mod sd_mod sbp2 loop snd_intel8x0 snd_ac97_codec snd_ac97_bus snd_pcm_oss snd_mixer_oss snd_pcm snd_timer tsdev snd serio_raw analog psmouse parport_pc parport usbhid rtc soundcore shpchp gameport floppy i2c_i801 snd_page_alloc pci_hotplug pcspkr intel_agp agpgart i2c_core evdev eth1394 ext3 jbd mbcache usb_storage scsi_mod ide_cd cdrom ide_disk uhci_hcd e100 mii ehci_hcd ohci1394 ieee1394 piix generic ide_core usbcore thermal processor fan CPU: 1 EIP: 0060:[<f8994b12>] Not tainted VLI EFLAGS: 00010246 (2.6.18-5-686 #1) EIP is at 0xf8994b12 eax: 00000000 ebx: c19d8000 ecx: dfc86ac0 edx: c180e6c0 esi: 00001000 edi: c1a13f74 ebp: b7fad000 esp: c1a13f60 ds: 007b es: 007b ss: 0068 Process hald-addon-keyb (pid: 2837, ti=c1a12000 task=dffc3000 task.ti=c1a12000) Stack: 00000000 dffc3000 c012d92d c1a13f6c c1a13f6c dfb6a240 f899ca3d b7fad000 00001000 c015a71c c1a13fa4 dfb6a240 fffffff7 00000001 c1a12000 c015ab68 c1a13fa4 00000000 00000000 00000000 00000004 0804e550 c0102c11 00000004 Call Trace: [<c012d92d>] autoremove_wake_function+0x0/0x2d [<f899ca3d>] evdev_read+0x0/0x164 [evdev] [<c015a71c>] vfs_read+0x9f/0x141 [<c015ab68>] sys_read+0x3c/0x63 [<c0102c11>] sysenter_past_esp+0x56/0x79 Code: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 <00> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 EIP: [<f8994b12>] 0xf8994b12 SS:ESP 0068:c1a13f60 * Maybe the above is already enough to make the system instable (at least there is no more output in dmesg), but I have tried to run microcode_ctl, too. There is two lines of output which I could not save. Basically it says that it successfully updated the microcode to some device node and something about the length of it. X crashes soon after, when I tried to log in. Reboot without further ado boots into a hanging system right away. I will attach the output of hwinfo, too, but I don't know what else I can provide for you. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User bwalle@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c29 Bernhard Walle changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |danielstefanmader@web.de --- Comment #29 from Bernhard Walle 2008-01-07 06:11:32 MST --- So, did I understand correctly that you get [<c012d92d>] autoremove_wake_function+0x0/0x2d [<f899ca3d>] evdev_read+0x0/0x164 [evdev] [<c015a71c>] vfs_read+0x9f/0x141 [<c015ab68>] sys_read+0x3c/0x63 [<c0102c11>] sysenter_past_esp+0x56/0x79 without even executing the microcode_ctl binary? Only after you loaded the kernel module? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User bwalle@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c31 --- Comment #31 from Bernhard Walle 2008-01-08 09:09:58 MST --- A colleague that I asked about the problem noted that the machine MCEs: Nov 5 17:01:49 chambertin kernel: MCE: The hardware reports a non fatal, correctable incident occurred on CPU 1. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User shaohua.li@intel.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c33 Shaohua Li changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |shaohua.li@intel.com --- Comment #33 from Shaohua Li 2008-01-08 20:49:37 MST ---

IA-32 Microcode Update Driver: v1.14a microcode: CPU0 updated from revision 0x21 to 0x2e, date = 08112004 microcode: CPU1 updated from revision 0x21 to 0x2e, date = 08112004 IA-32 Microcode Update Driver: v1.14a This suggests the microcode ctl is already called, ucode is already updated. version 0x2e is the latest ucode for CPU (0f-02-09). It's very strange a ucode update will cause CPU malfunctionl. Can you double check if the oops only occurs after microcode data is uploaded?

-- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User danielstefanmader@web.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c35 Daniel Mader changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|danielstefanmader@web.de | --- Comment #35 from Daniel Mader 2008-01-09 05:59:37 MST --- The installed Debian system behaves horrible ever since I played with microcode on it: I am no longer able to switch to a virtual terminal and test. I don't even know for sure if the microcode stuff is now activated or not during boot (it seems not, but still the system is odd). I am now trying with Knoppix 5.2 via ssh: * microcode is not loaded by default, so loading it! dmesg | tail -5 Mobile IPv6 eth1: no IPv6 routers present e100: eth1: e100_watchdog: link up, 100Mbps, full-duplex eth1: no IPv6 routers present IA-32 Microcode Update Driver: v1.14a * installing microcode.ctl: root@Knoppix:~# apt-get update [...] root@Knoppix:~# apt-get install microcode.ctl Paketlisten werden gelesen... Fertig Abh�ngigkeitsbaum wird aufgebaut... Fertig Die folgenden NEUEN Pakete werden installiert: microcode.ctl 0 aktualisiert, 1 neu installiert, 0 zu entfernen und 2489 nicht aktualisiert. Es m�ssen 21,6kB Archive geholt werden. Nach dem Auspacken werden 131kB Plattenplatz zus�tzlich benutzt. Hole:1 http://ftp.de.debian.org testing/contrib microcode.ctl 1.17-6 [21,6kB] Es wurden 21,6kB in 0s geholt (107kB/s) Vorkonfiguration der Pakete ... W�hle vormals abgew�hltes Paket microcode.ctl. (Lese Datenbank ... E: Sub-process /usr/bin/dpkg exited unexpectedly root@Knoppix:~# apt-get install microcode.ctl Paketlisten werden gelesen... Fertig Abh�ngigkeitsbaum wird aufgebaut... Fertig Die folgenden NEUEN Pakete werden installiert: microcode.ctl 0 aktualisiert, 1 neu installiert, 0 zu entfernen und 2489 nicht aktualisiert. Es m�ssen noch 0B von 21,6kB Archiven geholt werden. Nach dem Auspacken werden 131kB Plattenplatz zus�tzlich benutzt. Vorkonfiguration der Pakete ... W�hle vormals abgew�hltes Paket microcode.ctl. (Lese Datenbank ... 472420 Dateien und Verzeichnisse sind derzeit installiert.) Entpacke microcode.ctl (aus .../microcode.ctl_1.17-6_i386.deb) ... Richte microcode.ctl ein (1.17-6) ... Downloading a new version of microcode. microcode downloaded sucessfully Applying Intel IA32 Microcode update... done. /var/lib/dpkg/info/microcode.ctl.postinst: line 73: 4143 Speicherzugriffsfehler update-rc.d microcode.ctl start 80 S . >/dev/null dpkg: Fehler beim Bearbeiten von microcode.ctl (--configure): Unterprozess post-installation script mit Signal (Speicherzugriffsfehler) get�tet Fehler traten auf beim Bearbeiten von: microcode.ctl sh: line 1: 4146 Speicherzugriffsfehler /usr/bin/debsums --generate=nocheck -sp /var/cache/apt/archives E: Problem executing scripts DPkg::Post-Invoke 'if [ -x /usr/bin/debsums ]; then /usr/bin/debsums --generate=nocheck -sp /var/cache/apt/archives; fi' E: Sub-process returned an error code E: Sub-process /usr/bin/dpkg returned an error code (1) * and then pretty soon there is: Disconnecting: Corrupted MAC on input. * checking locally, the system is totally unstable (escpecially KDE/X) and dpgk/apt is not longer usable... * So concluding, I'd say that as long as microcode_ctl has not been called, everything is fine, but once called you can forget it... -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User danielstefanmader@web.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c37 --- Comment #37 from Daniel Mader 2008-01-09 07:12:38 MST --- Oops, sorry, didn't think about that. And the apt-get stuff is there because I wanted to document completely what I did. Hope it will help! -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 Bernhard Walle changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|bwalle@novell.com |shaohua.li@intel.com Status|NEEDINFO |NEW Info Provider|shaohua.li@intel.com | -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User shaohua.li@intel.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c40 --- Comment #40 from Shaohua Li 2008-01-10 00:54:31 MST --- Also can you to user a different version microcode data file, and see if there is any difference. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User shaohua.li@intel.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c42 --- Comment #42 from Shaohua Li 2008-01-10 01:05:34 MST --- BTW, you can get the old microcode data file in http://urbanmyth.org/microcode/ The ucode for the cpu is added in the microcode_ctl version of January 2006. So it's pretty old, I wonder why you didn't see it before. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User bwalle@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c44 --- Comment #44 from Bernhard Walle 2008-01-16 03:48:48 MST --- Of course! Download http://www.urbanmyth.org/microcode/ucode/intel-ia32microcode-29Aug2005.txt.b..., bunzip2 it, and replace your /etc/microcode.dat with that file. (Make a copy of that file first, maybe just /etc/microcode.dat.orig.) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User bwalle@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c46 --- Comment #46 from Bernhard Walle 2008-01-16 05:21:08 MST --- (In reply to comment #45 from Daniel Mader)

OK, I tried this on the openSUSE 10.3 system.

/var/log/messages after modprobe microcode microcode_ctl -u

So, I'd say, with that old microcode file, “updating” (because there's nothing to update because the file doesn't contain microcode for your CPU) works.

Jan 16 12:05:04 chambertin firmware.sh[5546]: Cannot find firmware file 'intel-ucode/0f-02-09' Jan 16 12:05:04 chambertin kernel: IA-32 Microcode Update Driver: v1.14a Jan 16 12:05:04 chambertin firmware.sh[5550]: Cannot find firmware file 'intel-ucode/0f-02-09'

That error message has nothing to do with the action microcode_ctl performs. That's the in-kernel firmware interface for which microcode files are not available. That's something for the future to replace the microcode_ctl userspace program at all. Shaohua, am I right? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User shaohua.li@intel.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c48 --- Comment #48 from Shaohua Li 2008-01-16 18:51:16 MST ---

Cannot find firmware file 'intel-ucode/0f-02-09' Yes, this doesn't matter.

But you should see something like:

microcode: CPU1 updated from revision 0x21 to 0x2e, date = 08112004 If not, then microcode update isn't updated. from comments #47, the hang is after you applied the update of http://www.urbanmyth.org/microcode/ucode/intel-ia32microcode-29Aug2005.txt.b..., right? If yes, either the microcode update driver is buggy or other code is buggy. I'll double check the driver. In the meantime, can you try the older version microcode update driver see if this is a regression.

-- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User danielstefanmader@web.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c50 Daniel Mader changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|danielstefanmader@web.de | --- Comment #50 from Daniel Mader 2008-01-31 03:02:40 MST --- Any news on the driver? And again: with slightly more detailed instructions I'd be happy to help... -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User danielstefanmader@web.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c52 --- Comment #52 from Daniel Mader 2008-02-05 12:00:22 MST --- Unfortunately, there is no microcode_2616.ko but only 2617 and up. I am using 2.6.22.16-0.1-default and the respective source. chambertin:~/microcode_build # make make -C /lib/modules/2.6.22.16-0.1-default/build M=/root/microcode_build make[1]: Entering directory `/usr/src/linux-2.6.22.16-0.1-obj/i386/default' make -C ../../../linux-2.6.22.16-0.1 O=../linux-2.6.22.16-0.1-obj/i386/default LD /root/microcode_build/built-in.o CC [M] /root/microcode_build/microcode_2620.o CC [M] /root/microcode_build/microcode_2622.o CC [M] /root/microcode_build/microcode_2621.o CC [M] /root/microcode_build/microcode_2617.o CC [M] /root/microcode_build/microcode_2618.o CC [M] /root/microcode_build/microcode_2619.o Building modules, stage 2. MODPOST 6 modules WARNING: "suspend_cpu_hotplug" [/root/microcode_build/microcode_2621.ko] undefined! CC /root/microcode_build/microcode_2617.mod.o LD [M] /root/microcode_build/microcode_2617.ko CC /root/microcode_build/microcode_2618.mod.o LD [M] /root/microcode_build/microcode_2618.ko CC /root/microcode_build/microcode_2619.mod.o LD [M] /root/microcode_build/microcode_2619.ko CC /root/microcode_build/microcode_2620.mod.o LD [M] /root/microcode_build/microcode_2620.ko CC /root/microcode_build/microcode_2621.mod.o LD [M] /root/microcode_build/microcode_2621.ko CC /root/microcode_build/microcode_2622.mod.o LD [M] /root/microcode_build/microcode_2622.ko make[1]: Leaving directory `/usr/src/linux-2.6.22.16-0.1-obj/i386/default' After inserting the module and calling microcode_ctl -u with intel-ia32microcode-29Aug2005.txt as /etc/microcode.dat, /var/log/messages says: Feb 5 19:52:54 chambertin kernel: IA-32 Microcode Update Driver: v1.14 Feb 5 19:54:22 chambertin kernel: microcode: collect_cpu_info : sig=0xf29, pf=0x4, rev=0x21 Feb 5 19:54:22 chambertin kernel: microcode: collect_cpu_info : sig=0xf29, pf=0x4, rev=0x21 Feb 5 19:54:22 chambertin kernel: Microcode Found. Feb 5 19:54:22 chambertin kernel: Header Revision 0x1 Feb 5 19:54:22 chambertin kernel: Loader Revision 0x1 Feb 5 19:54:22 chambertin kernel: Revision 0x2e Feb 5 19:54:22 chambertin kernel: Date 8/11/2004 Feb 5 19:54:22 chambertin kernel: Signature 0xf29 Feb 5 19:54:22 chambertin kernel: Type 0x0 Family 0xf Model 0x2 Stepping 0x9 Feb 5 19:54:22 chambertin kernel: Processor Flags 0x4 Feb 5 19:54:22 chambertin kernel: Checksum 0x7695d4dc Feb 5 19:54:22 chambertin kernel: microcode: CPU0 found a matching microcode update with revision 0x2e (current=0x21) Feb 5 19:54:22 chambertin kernel: Microcode Found. Feb 5 19:54:22 chambertin kernel: Header Revision 0x1 Feb 5 19:54:22 chambertin kernel: Loader Revision 0x1 Feb 5 19:54:22 chambertin kernel: Revision 0x2e Feb 5 19:54:22 chambertin kernel: Date 8/11/2004 Feb 5 19:54:22 chambertin kernel: Signature 0xf29 Feb 5 19:54:22 chambertin kernel: Type 0x0 Family 0xf Model 0x2 Stepping 0x9 Feb 5 19:54:22 chambertin kernel: Processor Flags 0x4 Feb 5 19:54:22 chambertin kernel: Checksum 0x7695d4dc Feb 5 19:54:22 chambertin kernel: microcode: CPU1 found a matching microcode update with revision 0x2e (current=0x21) Feb 5 19:54:22 chambertin kernel: microcode: CPU1 updated from revision 0x21 to 0x2e, date = 08112004 Feb 5 19:54:22 chambertin kernel: microcode: CPU0 updated from revision 0x21 to 0x2e, date = 08112004 The systems then soon hangs as usually... -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User danielstefanmader@web.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c54 --- Comment #54 from Daniel Mader 2008-02-05 13:59:38 MST --- the next lowest available: microcode_2617.ko -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User bwalle@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c56 --- Comment #56 from Bernhard Walle 2008-02-05 14:10:26 MST --- I have the same feeling. Maybe you can try a complete 2.6.17 kernel. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 Daniel Mader changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|permanent segfaults and freezes on P4 dual core |microcode.ctl causes segfaults and freezes on P4 |machine |machine -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User danielstefanmader@web.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c59 --- Comment #59 from Daniel Mader 2008-06-20 08:25:52 MDT --- Created an attachment (id=223409) --> (https://bugzilla.novell.com/attachment.cgi?id=223409) output of 'hwinfo > hwinfo.stdout' Yes, definitely an Intel P4. According to the datasheet, the system is as follows (please compare with the attached hwinfo output or the previously attached lspci output): * GIGABYTE 8IPE1000Pro Mainboard * Intel 865PE chipset with hyperthreading support * Intel Pentium 4 3000 "Northwood" -3000Mhz -512KB Cache OnDie -800Mhz bus -Hyper-Threading -FCPGA-478, 0.13µm -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User kent.liu@intel.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c61 Kent Liu changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|kent.liu@intel.com | --- Comment #61 from Kent Liu 2008-06-23 23:26:00 MDT --- We can't reproduce this in our lab, so we need to know if the issue is a regression from one settings to another settings. So what we want here is to find a distro that can work after loading microcode with microcode.ctl, so that we can compare the difference between them. Did you try a more recent Ubuntu, such as Ubuntu 8.04? It has new kernel and the microcode dat file is same as what OpenSuSE 11 has. If this can be reproduced there, we can confirm it's an common upstream issue. If this can't be reproduced, we should check if there is difference. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User kent.liu@intel.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c63 --- Comment #63 from Kent Liu 2008-07-06 17:49:04 MDT --- (In reply to comment #62 from Josua Dietze)

I have Debian Etch, Ubuntu 8.04, Fedora 9 beside openSUSE 11. All default kernels, but none of these others show the error; in fact, I managed to boot into openSUSE only by using the Ubuntu kernel. When you boot to Ubuntu and Fedora kernels, are you sure 'microcode_ctl -u' has been executed? If yes, are the microcode dat files same with openSuSE 11?

-- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User kent.liu@intel.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c65 --- Comment #65 from Kent Liu 2008-07-07 17:44:54 MDT --- Thanks! As same microcode data file causes difference appearance, the root cause may be not in microcode file, but in kernel or microcode.ctl utility. One possible try to isolate the kernel issue and microcode.ctl issue is to use F9 kernel on OpenSuSE and uses the microcode.ctl utility shipped in openSuSE. The kernel configuration should be same as what the kernel was used in F9 environment. By the way, I have a little suspection on -pae configuration in OpenSuSE kernel. If you can, please also test f9 or ubuntu 8.04 with -pae enabled. (or you can just copy the .config from OpenSuSE to f9/ubuntu and rebuild kernel). -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User digidietze@draisberghof.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c67 --- Comment #67 from Josua Dietze 2008-07-08 04:02:45 MDT --- O.K.; just give me two days or so (got to make a living too :-) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User digidietze@draisberghof.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c69 --- Comment #69 from Josua Dietze 2008-07-09 12:28:20 MDT --- First results ... On openSUSE 11: - Fedora kernel 2.6.25.9-76.fc9.i686 shows NO problems with microcode - openSUSE kernel 2.6.25.9-0.2-default shows NO problems with microcode On Fedora 9: - Fedora kernel 2.6.25.9-76.fc9.i686.PAE gives oops or reset when "microcode_ctl -u" Now I hope we're getting somewhere. I attached the log of the oops (via serial console). Sorry to say, my knowledge stops right there; I'm more of an application guy, you know, human readable code :-) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User digidietze@draisberghof.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c71 --- Comment #71 from Josua Dietze 2008-07-10 16:58:54 MDT --- Yes, this is reproduceable beyond doubt. Regarding this patch, you lost me somewhere on the way. I built the PAE kernel with the openSUSE config; then I inserted that "tlb_flush_all" command. But I had to manually add the "asm/tlbflush.h" reference, and still stage 2 bails out with an ominous "section mismatch" error. I suspect I should set some debug options ... Please assist me there ... -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User shaohua.li@intel.com added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c73 Shaohua Li changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #227136|0 |1 is obsolete| | --- Comment #73 from Shaohua Li 2008-07-10 19:36:11 MDT --- Created an attachment (id=227137) --> (https://bugzilla.novell.com/attachment.cgi?id=227137) debug patch Oops, last one is still buggy, try this one please. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User digidietze@draisberghof.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c75 --- Comment #75 from Josua Dietze 2008-07-11 10:04:57 MDT --- Well, I managed to get my patched kernel working with the serial console. If that's not what you are expecting I'll try the Bernhard Kernel ... Grub kernel line (normal mode just gave reset): "kernel /boot/vmlinuz-2.6.25.9-0.2-pae root=/dev/sda5 showopts ide=nodma apm=off acpi=off noresume nosmp noapic maxcpus=0 edd=off x11failsafe console=ttyS0,38400" Log: testbox:~ # modprobe -v microcode.dat insmod /lib/modules/2.6.25.9-0.2-pae/kernel/drivers/base/firmware_class.ko insmod /lib/modules/2.6.25.9-0.2-pae/kernel/arch/x86/kernel/microcode.ko testbox:~ # microcode_ctl -u microcode_ctl: wdouble fault, tss at c1809800 eip = c01e4100, esp = f7977fac eax = 00000021, ebx = c04ee620, ecx = 00000000, edx = c04eea1c esi = c04ee620, edi = f79780ec After that, hang ... If you mean HyperThreading by HT, I don't think my machine knows anything about that. At least the BIOS doesn't. From /proc/cpuinfo: processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 13 model name : Intel(R) Pentium(R) M processor 1.73GHz stepping : 8 cpu MHz : 1733.458 cache size : 2048 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss tm pbe nx up bts est tm2 bogomips : 3469.22 clflush size : 64 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User danielstefanmader@web.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c77 --- Comment #77 from Daniel Mader 2008-07-14 02:48:36 MDT --- Hi, I've been watching this awe-inspiring thread and I thought the workaround for me is, to switch from kernel-pae to kernel-default on our system (i.e. the original system where I hit this bug). Unfortunately this did not change anything, the system segfaults and freezes like it ever did. So at least here it is not a matter of pae- or default-flavor (and I don't remember a pae-version for 10.3 at all). BTW: what triggers the installation of pae vs. default? All systems which get pae by default hav no more than 1.5GB RAM... -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=336277 User digidietze@draisberghof.de added comment https://bugzilla.novell.com/show_bug.cgi?id=336277#c79 --- Comment #79 from Josua Dietze 2008-07-16 10:44:51 MDT --- (In reply to comment #76 from Shaohua Li)

Hi Josua, looks this is a different system. The original cpu family/modle/stepping is f/2/9, the system you test is 6/13/8.

What does that mean? Do I have to open a new bug? openSUSE installed the PAE kernel as a default, so there is clearly something to improve. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.