https://bugzilla.novell.com/show_bug.cgi?id=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c5 Carlos Robinson changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|robin.listas@telefonica.net | --- Comment #5 from Carlos Robinson 2008-01-07 18:01:41 MST --- Created an attachment (id=189654) --> (https://bugzilla.novell.com/attachment.cgi?id=189654) requested info Reply to Comment #3 From Oliver Neukum

Could you add this line:

cat /proc/interrupts >> /home/cer/marca.log

Done. Today of all days it did not want to cooperate being lazy while I was looking: I guess it knew you were waiting and behaved naughtily. I had to stop mail fetch, ntp daemon, and wait for a minute in several tries, but at last I got it. Search for the time mark "=====> 01:35:15" in the attached file: it jumps to "01:35:43", almost half a minute stopped. There are a few smaller jumps, like "01:32:46..48", "01:33:55..57"; ah, another big one: 01:34:31..01:35:06". A one second jump could be doubtful, but a half a minute one I suppose not. I don't know what you are looking for, but the timer interrupt seems to be ticking :-? -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c7 --- Comment #7 from Carlos Robinson 2008-01-08 16:59:44 MST --- Ah! How can that be? If they are still 19.2 ticks/second, I should have had 537 ticks. I don't know how the timer works in linux, but I assume that, as the clock shows the correct time, the system uses TSC to estimate the lost interrupts correctly. Is that so? Now, questions... can this be a hardware problem and the timer chip doesn't produces the interrupts? However, in the test the other day with 10.2 there were no symptoms. Is the kernel reprogramming the chip to go slower? Is the kernel ignoring the interrupts? The only thing I can do, as I lack the knowledge of kernel internals, is to try again in 10.2 and measure the ticks over a period. Do you want me to try that? Note 1: This thing must be related to Bug 350981 Note 2: These days I'm using my personalized kernel, instead of the default suse kernel I used the other day for testing (which has the same symptoms). I play with different settings in an attempt to solve this. These are my different settings (diff -y --suppress-common-lines .config /boot/config-2.6.22.13-0.3-default): mine yours # Linux kernel version: 2.6.22.13 < # Sun Dec 30 14:30:44 2007 < CONFIG_BROKEN_ON_SMP=y | CONFIG_LOCK_KERNEL=y CONFIG_LOCALVERSION="-cer" | CONFIG_LOCALVERSION="-default" > CONFIG_CPUSETS=y > CONFIG_STOP_MACHINE=y CONFIG_HIGH_RES_TIMERS=y | # CONFIG_HIGH_RES_TIMERS is not set # CONFIG_SMP is not set | CONFIG_SMP=y # CONFIG_PARAVIRT is not set | CONFIG_PARAVIRT=y > CONFIG_VMI=y # CONFIG_M586 is not set | CONFIG_M586=y CONFIG_MPENTIUM4=y | # CONFIG_MPENTIUM4 is not set > CONFIG_X86_PPRO_FENCE=y > CONFIG_X86_F00F_BUG=y CONFIG_X86_GOOD_APIC=y | CONFIG_X86_ALIGNMENT_16=y CONFIG_X86_USE_PPRO_CHECKSUM=y < CONFIG_X86_TSC=y < CONFIG_X86_CMOV=y < > CONFIG_NR_CPUS=32 > CONFIG_SCHED_SMT=y > CONFIG_SCHED_MC=y > # CONFIG_PREEMPT_BKL is not set > CONFIG_X86_MCE_P4THERMAL=y # CONFIG_TOSHIBA is not set | CONFIG_TOSHIBA=m # CONFIG_I8K is not set | CONFIG_I8K=m > # CONFIG_IRQBALANCE is not set # CONFIG_HZ_250 is not set | CONFIG_HZ_250=y CONFIG_HZ_300=y | # CONFIG_HZ_300 is not set CONFIG_HZ=300 | CONFIG_HZ=250 > CONFIG_HOTPLUG_CPU=y > CONFIG_SUSPEND_SMP=y > CONFIG_ACPI_HOTPLUG_CPU=y # CONFIG_DMASCC is not set < # CONFIG_IRPORT_SIR is not set < # CONFIG_SBPCD is not set < # CONFIG_CM206 is not set < # CONFIG_CDU31A is not set < # CONFIG_NI5010 is not set < # CONFIG_PCMCIA_XIRTULIP is not set < # CONFIG_ISDN_DRV_LOOP is not set < # CONFIG_HYSDN is not set < # CONFIG_RISCOM8 is not set < # CONFIG_STALLION is not set < # CONFIG_ISTALLION is not set < # CONFIG_I2C_ELEKTOR is not set < # CONFIG_DEBUG_INFO is not set | CONFIG_DEBUG_INFO=y CONFIG_DEBUG_RODATA=y | # CONFIG_DEBUG_RODATA is not set > CONFIG_GENERIC_PENDING_IRQ=y > CONFIG_X86_SMP=y > CONFIG_X86_HT=y > CONFIG_X86_TRAMPOLINE=y -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c9 --- Comment #9 from Carlos Robinson 2008-01-08 17:45:37 MST --- One of my tests was to recompile with NO_HZ last December 25th: no effect. But if you want to be sure, I'll reboot with the default kernel with that option. cpufreq=no --> http://lists.opensuse.org/opensuse/2007-12/msg02136.html (no effect) http://lists.opensuse.org/opensuse/2007-12/msg02754.html kernel recompiled with NO_HZ disabled: no effect. Ok, I edited my grub menu.lst and will reboot now. I will report back. -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c11 Carlos Robinson changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|robin.listas@telefonica.net | --- Comment #11 from Carlos Robinson 2008-01-08 20:38:45 MST --- Worse luck. :-( I sent the computer to hibernate to disk, as I always do, and it crashed (keyboard stuck), about 1.5 centimeters in the progress bar: and that's all I could see, as there is no verbose mode for hibernating, except that damm cute green display with a progress bar that doesn't allow us "hackers" to see what is going on, or not going on, as this time. First time in one or two months. I hesitate to restart it again in nohz=off mode, I hope you understand that. I had also a warning during boot in this mode: first time it stopped during the network start phase, and I rebooted it (I thought the adsl might be down just at that moment). The second time it slowed for a few seconds there, then went on, so I didn't give it much thought. But seeing that it crashed/locked later when attempting to hibernate, all I can say is that the nohz=off mode causes my computer to be unstable. I don't remember it to be unstable when I selected this option at kernel compile time, though... And now I'm really off to sleep. You don't see it in the bugzilla log, but here it is 4:35 AM. I just rebooted to write this for you now instead of tomorrow. -- 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=350980 User oneukum@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c12 Oliver Neukum changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |robin.listas@telefonica.net --- Comment #12 from Oliver Neukum 2008-01-21 02:32:07 MST --- Time to break out the sledgehammer. Could you try with "noapic" added to the kernel command line? -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c14 Carlos Robinson changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|robin.listas@telefonica.net | --- Comment #14 from Carlos Robinson 2008-01-21 05:03:59 MST --- Ok, I'm now in "noapic" mode. I left the computer running in runlevel 3, router powered off, with my script detecting pauses, and none was felt - except a curious 2" pause about every minute: =====> 12:04:50 (******* 2 *******) =====> 12:05:50 (******* 2 *******) =====> 12:06:52 (******* 2 *******) =====> 12:07:54 (******* 2 *******) =====> 12:08:56 (******* 2 *******) =====> 12:09:58 (******* 2 *******) =====> 12:10:59 (******* 2 *******) =====> 12:12:01 (******* 2 *******) =====> 12:13:02 (******* 2 *******) I don't feel the computer slow, contrary to what I though. I have saved the output of dmesg and the /var/log/boot.msg in case you want them. I will try now to hibernate. -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c16 --- Comment #16 from Carlos Robinson 2008-02-02 05:04:24 MST --- I have tried the new kernel: vmlinuz-2.6.22.16-0.1-default. At first I thought things were better, but no, the problem remains. Then I recompiled the kernel with the "tickless" option: # CONFIG_NO_HZ is not set which I assume is more or less equivalent to the "nohz=off" in the startup line; however, the computer gets lazy, as usual (ie, the problem remains), but, hybernate works (with "nohz=off" it crashes). This is the kernel version I have currently active. Go figure. This is a log of "laziness" with my kernel during normal use: =====> 19:08:43 (******* 2 *******) =====> 19:08:57 (******* 5 *******) =====> 19:09:21 (******* 14 ***************************) =====> 19:09:58 (******* 34 ***************************) =====> 19:10:35 (******* 28 ***************************) =====> 19:10:41 (******* 2 *******) =====> 19:11:13 (******* 30 ***************************) =====> 19:11:51 (******* 34 ***************************) =====> 19:11:52 (******* 2 *******) =====> 19:12:28 (******* 36 ***************************) =====> 19:13:05 (******* 36 ***************************) =====> 19:13:43 (******* 37 ***************************) =====> 19:14:20 (******* 36 ***************************) =====> 19:14:52 (******* 31 ***************************) -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c18 --- Comment #18 from Carlos Robinson 2008-02-04 06:36:58 MST --- (In reply to comment #17 from Oliver Neukum)

Can you try that kernel with your current config options with

a) noapic b) irqfixup

in the kernel command line?

Is that two different tests, or one test with both parameters? As to "noapic" I already did, see #15. No good. Do you expect different results with this kernel? I will try "irqfixup" perhaps tonight. -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c20 Carlos Robinson changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|robin.listas@telefonica.net | --- Comment #20 from Carlos Robinson 2008-02-05 06:15:32 MST ---

a) noapic

The boot "feels" slow. No "laziness" in the brief test (brief because I didn't go for a coffee to let it time to be lazy). Hibernate works. It does work better. There is this imprecise feeling of being slow, though. I'd have to use a chronometer to time the boot to make sure.

b) irqfixup

I used "apic irqfixup" The boot stopped for ±30" on the line: Starting udevd udevd-event[1691]: node_symlink: device node '/dev/rtc0' already exists, link '/dev/rtc' will not overwrite it No idea why. Using a test user (cer2) pressing the power off button did not offer the option to hibernate. The normal user (cer) is hibernated, no questions asked (same thing for test "a"). Saw this in the log - it's new to me: Feb 5 12:33:19 nimrodel syslog-ng[3811]: last message repeated 4 times Feb 5 12:33:19 nimrodel gdm[7475]: GLib-CRITICAL: g_key_file_get_string: assertion `key_file != NULL' failed Feb 5 12:33:19 nimrodel gdm[7475]: GLib-CRITICAL: g_key_file_get_string: assertion `key_file != NULL' failed Feb 5 12:33:19 nimrodel gdm[7475]: GLib-CRITICAL: g_key_file_free: assertion `key_file != NULL' failed The computer felt ok, no laziness. Then I went for a coffee, and then, yes, I saw a 37" pause at 12:52:35 =====> 12:51:58 (1) CPU0 0: 91435 IO-APIC-edge timer 1: 2452 IO-APIC-edge i8042 6: 4 IO-APIC-edge floppy 7: 0 IO-APIC-edge parport0 8: 12 IO-APIC-edge rtc 9: 6 IO-APIC-fasteoi acpi 10: 2 IO-APIC-edge MPU401 UART 12: 18436 IO-APIC-edge i8042 14: 30660 IO-APIC-edge ide0 15: 82384 IO-APIC-edge ide1 16: 1751 IO-APIC-fasteoi ehci_hcd:usb1, eth0 17: 0 IO-APIC-fasteoi uhci_hcd:usb2 18: 0 IO-APIC-fasteoi uhci_hcd:usb3 19: 1780 IO-APIC-fasteoi Intel 82801BA-ICH2 NMI: 0 LOC: 126068 ERR: 0 MIS: 0 =====> 12:52:35 (******* 37 ***************************) CPU0 0: 91468 IO-APIC-edge timer 1: 2452 IO-APIC-edge i8042 6: 4 IO-APIC-edge floppy 7: 0 IO-APIC-edge parport0 8: 12 IO-APIC-edge rtc 9: 6 IO-APIC-fasteoi acpi 10: 2 IO-APIC-edge MPU401 UART 12: 18436 IO-APIC-edge i8042 14: 30668 IO-APIC-edge ide0 15: 82425 IO-APIC-edge ide1 16: 1751 IO-APIC-fasteoi ehci_hcd:usb1, eth0 17: 0 IO-APIC-fasteoi uhci_hcd:usb2 18: 0 IO-APIC-fasteoi uhci_hcd:usb3 19: 1780 IO-APIC-fasteoi Intel 82801BA-ICH2 NMI: 0 LOC: 126110 ERR: 0 MIS: 0 I have the boot logs saved as /var/log/boot_test_apic_irqfixup.msg /var/log/boot_test_noapic.msg, if you want them I'll attach them. -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c22 --- Comment #22 from Carlos Robinson 2008-02-06 12:02:32 MST --- Created an attachment (id=193458) --> (https://bugzilla.novell.com/attachment.cgi?id=193458) Test number 5, noapic -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c24 --- Comment #24 from Carlos Robinson 2008-02-06 12:11:45 MST --- Created an attachment (id=193464) --> (https://bugzilla.novell.com/attachment.cgi?id=193464) test (6) with apic irqfixup active -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c26 Carlos Robinson changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|robin.listas@telefonica.net | --- Comment #26 from Carlos Robinson 2008-02-06 12:18:19 MST --- Created an attachment (id=193467) --> (https://bugzilla.novell.com/attachment.cgi?id=193467) test 6, boot log (irqfixup) Attached go the boot logs of both tests, which have the real kernel command line used. -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c29 --- Comment #29 from Carlos Robinson 2008-02-07 03:51:55 MST --- (In reply to comment #27 from Oliver Neukum)

That shows that your box loses interrupts when the APIC is used.

Only when I use 10.3. Remember that I tested 10.2 from my other partition and it runs OK. That proves that it is the 10.3 kernel who is bad, not my hardware. https://bugzilla.novell.com/show_bug.cgi?id=350981#c4 -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c31 --- Comment #31 from Carlos Robinson 2008-02-07 06:26:10 MST --- Created an attachment (id=193604) --> (https://bugzilla.novell.com/attachment.cgi?id=193604) marker log inside 10.2 with apic -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c33 --- Comment #33 from Carlos Robinson 2008-02-07 06:28:28 MST --- Created an attachment (id=193606) --> (https://bugzilla.novell.com/attachment.cgi?id=193606) Boot log of 10.2 session -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c35 --- Comment #35 from Carlos Robinson 2008-02-08 03:56:39 MST --- Ah, ok. However, it is curious that both things fail, and that few people are affected. Couldn't the timer go wrong because the kernel looses interrupts? I really don't know how the timer works nowdays, and even less in Linux; the only one I know about is the old ticker interrupt at about 17 times per second. But it is a curious coincidence. -- 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=350980 User stefan.fent@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c37 --- Comment #37 from Stefan Fent 2008-05-13 05:26:53 MST --- It would be nice to know which system we're talking about, so could you please add the output of lspci -nn? -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c39 --- Comment #39 from Carlos Robinson 2008-05-13 06:26:54 MST --- Created an attachment (id=214761) --> (https://bugzilla.novell.com/attachment.cgi?id=214761) lspci -nn output -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c41 --- Comment #41 from Carlos Robinson 2008-05-13 06:39:10 MST --- (In reply to comment #37 from Stefan Fent)

It would be nice to know which system we're talking about, so could you please add the output of lspci -nn?

Done (In reply to comment #38 from Thomas Renninger)

And cat /proc/cpuinfo

Done. I have detected on the mail list somebody with possibly the same problem: http://lists.opensuse.org/opensuse/2008-05/msg00293.html I told him to report here but he hasn't replied. I have insisted (a minute ago) again by direct mail, it's all I can do. But at least you can see that there are more people out there with this problem, only that they don't report. I suppose that's typical. -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c43 --- Comment #43 from Carlos Robinson 2008-05-13 07:11:39 MST --- (In reply to comment #42 from Thomas Renninger)

Better try with 11.0. This probably came in with the clockevents/no_hz patches. There has quite some work been done meanwhile. It's not worth looking at for 10.3. Please reopen if you still see this problem in 11.0. If this is still in 11.0, this could even get a bit higher priority...

I want to, but I can't. For two weekends I haven't been able to upgrade factory because the OSS repo was broken. Now that's been solved, but it is a 1.4 GiB download that takes a minimum of 4 hours to do, actually 8 hours, meaning that the repo has changed again while I try to update... so, hopefully, I will not be able to upgrade my factory till saterday afternoon. Hopefully. There are several tests that I want to do in factory and I can't. -- 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=350980 Carlos Robinson changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #214762|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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c45 --- Comment #45 from Carlos Robinson 2008-05-16 05:13:45 MST --- (In reply to comment #42 from Thomas Renninger)

Better try with 11.0. This probably came in with the clockevents/no_hz patches. There has quite some work been done meanwhile. It's not worth looking at for 10.3. Please reopen if you still see this problem in 11.0. If this is still in 11.0, this could even get a bit higher priority...

Finally, I have been able to update factory, and I have run a test. I unplugged the network, left my marker script running, and went away for a cup of tea and a stroll. Twenty minutes later, I detected no "laziness". The screen saver had kicked in, beagle had finished, and the script didn't appear to have logged any problem. So far so good... I can't certify that this bug is finally solved, but it certainly looks that way. Nice! :-) -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c47 Carlos Robinson changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |REOPENED Component|Kernel |Kernel OS/Version|openSUSE 10.3 |openSUSE 11.0 Product|openSUSE 10.3 |openSUSE 11.0 Resolution|WONTFIX | --- Comment #47 from Carlos Robinson 2008-11-23 13:25:17 MST --- Sorry, chaps... it's back! :-( The symptoms are back in 11.0 with kernel 2.6.25.18-0.2-pae - it did not happen till the kernel was updated. Linux version 2.6.25.18-0.2-pae (geeko@buildhost) (gcc version 4.3.1 20080507 (prerelease) [gcc-4_3-branch revision 135036] (SUSE Linux) ) #1 SMP 2008-10-21 16:30:26 +0200 I didn't notice it earlier because the network was active (network interrupts keep this problem from hapening). That job finished, network is quiet, and the symptoms came back. Quite rare occurrences, but I have had a few. This is what called my attention on the "warn" log (yes, I keep a terminal tailing that log): Nov 23 13:59:55 nimrodel upsmon[11226]: Poll UPS [myups@localhost] failed - Data stale Nov 23 13:59:55 nimrodel named[3192]: *** POKED TIMER *** Nov 23 13:59:55 nimrodel spamd[484]: prefork: sysread(10) failed after 300 secs at /usr/lib/perl5/vendor_perl/5.10.0/Mail/SpamAssassin/SpamdF orkScaling.pm line 648. Nov 23 13:59:55 nimrodel spamd[514]: prefork: sysread(11) failed after 300 secs at /usr/lib/perl5/vendor_perl/5.10.0/Mail/SpamAssassin/SpamdF orkScaling.pm line 648. Nov 23 13:59:55 nimrodel spamd[11694]: Use of uninitialized value $selerr in concatenation (.) or string at /usr/lib/perl5/vendor_perl/5.10.0 /Mail/SpamAssassin/SpamdForkScaling.pm line 329. Nov 23 13:59:56 nimrodel spamd[11694]: prefork: select returned error on server filehandle: And sure enough, at that hour I had in my "marker" script output a hiccup: =====> 13:59:56.051383857 (******* 1453 ***************************) machine paused for a full 1453 seconds while I lunched! No, it was not an hybernation pause, those are logged in my system. Interrupts were lost, of course: =====> 13:35:43.415850755 (1) CPU0 0: 29297542 IO-APIC-edge timer 1: 332079 IO-APIC-edge i8042 3: 8 IO-APIC-edge 4: 8 IO-APIC-edge 6: 7 IO-APIC-edge floppy 8: 0 IO-APIC-edge rtc0 9: 0 IO-APIC-fasteoi acpi 10: 9 IO-APIC-edge MPU401 UART 12: 1292007 IO-APIC-edge i8042 14: 1887391 IO-APIC-edge ide0 15: 2101920 IO-APIC-edge ide1 16: 1530281 IO-APIC-fasteoi ehci_hcd:usb2, eth0 17: 779636 IO-APIC-fasteoi Intel 82801BA-ICH2 18: 0 IO-APIC-fasteoi uhci_hcd:usb3 19: 471397 IO-APIC-fasteoi uhci_hcd:usb1, uhci_hcd:usb4 NMI: 0 Non-maskable interrupts LOC: 13316637 Local timer interrupts RES: 0 Rescheduling interrupts CAL: 0 function call interrupts TLB: 0 TLB shootdowns TRM: 0 Thermal event interrupts SPU: 0 Spurious interrupts ERR: 0 MIS: 0 =====> 13:59:56.051383857 (******* 1453 ***************************) CPU0 0: 29297656 IO-APIC-edge timer 1: 332079 IO-APIC-edge i8042 3: 8 IO-APIC-edge 4: 8 IO-APIC-edge 6: 7 IO-APIC-edge floppy 8: 0 IO-APIC-edge rtc0 9: 0 IO-APIC-fasteoi acpi 10: 9 IO-APIC-edge MPU401 UART 12: 1292007 IO-APIC-edge i8042 14: 1887404 IO-APIC-edge ide0 15: 2101966 IO-APIC-edge ide1 16: 1530282 IO-APIC-fasteoi ehci_hcd:usb2, eth0 17: 779636 IO-APIC-fasteoi Intel 82801BA-ICH2 18: 0 IO-APIC-fasteoi uhci_hcd:usb3 19: 471420 IO-APIC-fasteoi uhci_hcd:usb1, uhci_hcd:usb4 NMI: 0 Non-maskable interrupts LOC: 13316723 Local timer interrupts RES: 0 Rescheduling interrupts CAL: 0 function call interrupts TLB: 0 TLB shootdowns TRM: 0 Thermal event interrupts SPU: 0 Spurious interrupts ERR: 0 MIS: 0 =====> 13:59:57.088127523 (1 I have more pauses logged. Often they last 10..20 seconds. The system clocksource is currently tsc. Any more info you want? Hopefully you can track it to some change done on the last kernel update. And no, I'm afraid I can't test in factory, it hard locks on me. Sorry. Old related bugs: 350981, 344356. -- 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=350980 User trenn@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c48 Thomas Renninger changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO Info Provider| |robin.listas@telefonica.net --- Comment #48 from Thomas Renninger 2008-11-24 07:55:41 MST --- Have you already searched for a BIOS update? Please also attach dmidecode and best also hwinfo to get an even better idea what machine we have. -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c50 --- Comment #50 from Carlos Robinson 2008-11-24 11:30:28 MST --- Created an attachment (id=254913) --> (https://bugzilla.novell.com/attachment.cgi?id=254913) dmidecode requested -- 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=350980 User trenn@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c52 Thomas Renninger changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED Priority|P5 - None |P3 - Medium --- Comment #52 from Thomas Renninger 2008-11-25 05:11:50 MST ---

It happens only with certain suse versions Ok, I think I got it, sorry about that.

Let's try to limit this to as less patches as possible, then try out the ones that look guilty. This 11.0 kernel was the first which did not work: 2.6.25.18-0.2-pae Did you do upates before, can you remember the last working kernel? For 10.3 all kernels (you tested) were broken. 10.2 and before worked well. This is all about i386-pae kernels or did you also try x86_64 or i386 without -pae support kernels? -- 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=350980 User trenn@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c54 Thomas Renninger changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |robin.listas@telefonica.net --- Comment #54 from Thomas Renninger 2009-01-30 08:32:37 MST --- Sorry, I lost track of this one. Can you give 11.1 a try, please. This might have been fixed meanwhile. -- 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=350980 User robin.listas@telefonica.net added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c56 --- Comment #56 from Carlos Robinson 2009-02-10 07:22:17 MST --- Bug 448007 counts as fixed, but the fix means I have to use an experimental kernel: kernel-debug-2.6.27.13-SL111_BRANCH_20090203134609_5784a3e1.i586.rpm There is no info on whether this fix will be propagated to standard 11.1, which means that, with the info I have, I'll have to skip upgrading to 11.1, even though I have your complimentary DVD (and torch!) for my collaboration as translator and tester. So, if you want me to test this bug with the above experimental kernel (in my test partition), I'll try. If not, we'll have to wait till 11.2. And no, before you ask: I can not install factory, because that's the 11.1 test partition. -- 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=350980 User trenn@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=350980#c58 --- Comment #58 from Thomas Renninger 2009-02-10 08:45:55 MST --- To be clear: do not take the debug, but -default + -default-base kernel packages from the SLE11-BRANCH/arch directory. -- 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.