[Bug 506790] New: Machine hangs after powernow fix with latest Kernel
http://bugzilla.novell.com/show_bug.cgi?id=506790 Summary: Machine hangs after powernow fix with latest Kernel Classification: openSUSE Product: openSUSE 11.1 Version: Final Platform: x86-64 OS/Version: Other Status: NEW Severity: Major Priority: P5 - None Component: Kernel AssignedTo: bnc-team-screening@forge.provo.novell.com ReportedBy: mail@cynapses.org QAContact: qa@suse.de Found By: Community User Since I'm running kernel-default-2.6.27.21-0.1.2 my server hangs from time to time. The CPU is a AMD Athlon(tm) 64 Processor 3800+. Let me know if you need more information about the system. The latest messages are: 7843 May 24 17:30:27 milliways kernel: powernow-k8: error - out of sync, fix 0x10 0xc, vid 0x8 0x8 7844 May 24 17:31:17 milliways kernel: powernow-k8: error - out of sync, fix 0x10 0xc, vid 0x8 0x8 7845 May 24 17:31:21 milliways kernel: powernow-k8: error - out of sync, fix 0xa 0x10, vid 0xc 0xc 7846 May 24 17:31:21 milliways kernel: powernow-k8: ph2 null fid transition 0x10 7847 May 24 17:31:27 milliways kernel: powernow-k8: error - out of sync, fix 0x10 0xc, vid 0x8 0x8 7848 May 24 17:31:30 milliways kernel: powernow-k8: error - out of sync, fix 0xa 0x10, vid 0xc 0xc 7849 May 24 17:31:45 milliways kernel: powernow-k8: error - out of sync, fix 0x10 0xc, vid 0x8 0x8 7850 May 24 17:31:48 milliways kernel: powernow-k8: error - out of sync, fix 0xc 0x10, vid 0xa 0xa 7851 May 25 10:47:03 milliways syslog-ng[2355]: syslog-ng starting up; version='2.0.9' -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=506790 User mail@cynapses.org added comment http://bugzilla.novell.com/show_bug.cgi?id=506790#c1 --- Comment #1 from Andreas Schneider <mail@cynapses.org> 2009-05-25 09:58:30 MDT --- This happened again in 24h now. I hope unloading the powernow_k8 module helps as a workaround. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=506790 User mail@cynapses.org added comment http://bugzilla.novell.com/show_bug.cgi?id=506790#c2 Andreas Schneider <mail@cynapses.org> changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|bnc-team-screening@forge.pr |trenn@novell.com |ovo.novell.com | --- Comment #2 from Andreas Schneider <mail@cynapses.org> 2009-05-25 10:01:44 MDT --- * Sat Jan 24 2009 trenn@suse.de - patches.fixes/x86_cpufreq_powernow-k8_acpi_latency_values.patch: X86 powernow-k8 cpufreq: Get transition latency from acpi _PSS object (bnc#464461). Could this patch bite me? -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=506790 User mail@cynapses.org added comment http://bugzilla.novell.com/show_bug.cgi?id=506790#c3 --- Comment #3 from Andreas Schneider <mail@cynapses.org> 2009-05-25 10:11:26 MDT --- Created an attachment (id=294260) --> (http://bugzilla.novell.com/attachment.cgi?id=294260) acpidump of the machine -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=506790 User mail@cynapses.org added comment http://bugzilla.novell.com/show_bug.cgi?id=506790#c4 --- Comment #4 from Andreas Schneider <mail@cynapses.org> 2009-06-02 03:34:34 MDT --- The machine runs just fine without the powernow_k8 module. If I load it the machine freezes. Any progress here? -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=506790 User trenn@novell.com added comment http://bugzilla.novell.com/show_bug.cgi?id=506790#c5 Thomas Renninger <trenn@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO CC| |mark.langsdorf@amd.com Info Provider| |mail@cynapses.org --- Comment #5 from Thomas Renninger <trenn@novell.com> 2009-06-03 06:38:49 MDT --- This is really weird. There is no cpufreq information in the ACPI BIOS tables. So about comment #2: No the patch should not affect you (unless there is a really big hammer of bug in it). This is a single core system? Then the information seem to come from a deprecated/old PSB table. I cannot see any printk which could verify this in syslog, though. Maybe a warning/notice should be added if PSB table is used... Did this ever work (powernow-k8 driver) on this machine? What does: cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_transition_latency tell us? Adding Mark, he might be interested or could tell us more about PSB and whether it could/should get used at all. If this never worked or powernow-k8 never got tested on this machine, then better upgrade the BIOS, otherwise I'd like to dig a bit more and find the regression. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=506790 Thomas Renninger <trenn@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Priority|P5 - None |P3 - Medium -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=506790 User mail@cynapses.org added comment http://bugzilla.novell.com/show_bug.cgi?id=506790#c6 --- Comment #6 from Andreas Schneider <mail@cynapses.org> 2009-06-03 06:45:58 MDT --- Yes, this is a single core system, AMD Athlon(tm) 64 Processor 3800+, 2400 MHz. milliways:~ # modprobe powernow_k8 milliways:~ # cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_transition_latency 1240000 I'm not able to do a BIOS update, this is a server at a ISP. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=506790 Andreas Schneider <mail@cynapses.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|mail@cynapses.org | -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=506790 User trenn@novell.com added comment http://bugzilla.novell.com/show_bug.cgi?id=506790#c7 Thomas Renninger <trenn@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WONTFIX --- Comment #7 from Thomas Renninger <trenn@novell.com> 2009-06-03 07:20:06 MDT --- The transition latency of: 1240000 corresponds to the wrongly calculated one if no ACPI table is used. This will result in an ondemand governor sampling rate of 1,24 seconds which makes PowerNow! rather useless or say will introduce a sever performance regression without manually lowering the ondemand governor sampling rate. On the other hand, this high sampling rate should be save and not the reason for powernow-k8 driver breaking/freezing the machine. I expect this comes from wrong PSB table values which I am not willing to debug any further. I also expect powernow-k8 driver never worked on this machine and this is not a regression. Please reopen if powernow-k8 already worked. Blacklisting powernow-k8 for this machine is IMO the right thing to do for this BIOS. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com