https://bugzilla.novell.com/show_bug.cgi?id=387207
User trenn@novell.com added comment
https://bugzilla.novell.com/show_bug.cgi?id=387207#c35
Thomas Renninger
From the huge difference, I expect we have a CPU problem here. Possibly graphics cards, but I very much expect the CPU.
The first report states a Pentium 4 in the description, so it cannot be cpufreq. It looks like the processor sleep states (C-states) are not entered correctly. The Pentium 4 probably has no deeper sleep states (only C1), it could be that something is still spinning instead of entering the idle function, like if idle=poll is passed? Can you do: echo 1 > /proc/sys/kernel/sysrq Then push the SysRq button (near Scroll/Lock button) and also "P" Do this when the machine is idle. You should then see a backtrace at the end of: dmesg |less It should look like: ----------------- RIP: 0010:[<ffffffff80221970>] [<ffffffff80221970>] native_safe_halt+0x6/0x8 .. Call Trace: Inexact backtrace: [<ffffffff8020b115>] ? default_idle+0x43/0x78 [<ffffffff8020b0d2>] ? default_idle+0x0/0x78 [<ffffffff8020b08a>] ? cpu_idle+0x92/0xda [<ffffffff804447fc>] ? start_secondary+0x408/0x417 ----------------- Important is the RIP line and the backtrace (like above). If you are not in the default_idle function, try again some more times when no process is running. Can you attach the stripped output like above. The C-states statistics are exported to userspace via: /proc/acpi/processor/*/power Please also run powertop and tell us if you see any noticable message/warning. -- 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.