Created attachment 383184 [details] dmesg on bootup (no suspending) Description of problem: Each time laptop is resumed from suspend, I observe the following behavior: 1. cpu core temperatures rapidly (about 1 minute) steadily soar up on idle machine from around 50C past 80C, heating stops only by reboot or poweroff. 2. Two other temperatures are falsely reported as constantly 27C: * /proc/acpi/thermal_zone/TZ01/temperature * sensors -> acpitz-virtual-0 (AKA libsensors temp1) Version-Release number of selected component (if applicable): DVD install of x86_64 on This has been a problem with Fedora 11 if not earlier and continues to be a problem with Fedora 12. How reproducible: Every time machine is resumed from suspend. Steps to Reproduce: 1. Suspend 2. Resume 3. Watch cpus bake Actual results: Expected results: Idle machine should not threaten to melt its CPUs. Additional info: Laptop specs at www.newageinc.com/pdfs/CPQ%20A916NR.pdf Compaq A916NR Presario Intel Pentium Dual-Core Mobile Processor T2330 More information available on request. What would be helpful?
Created attachment 383189 [details] dmesg after resume. CPU core temps are drifting up. Note the temp1 below stays constant; the others drift higher and higher until reboot. $ sensors acpitz-virtual-0 Adapter: Virtual device temp1: +26.8°C (crit = +100.0°C) coretemp-isa-0000 Adapter: ISA adapter Core 0: +66.0°C (high = +100.0°C, crit = +100.0°C) coretemp-isa-0001 Adapter: ISA adapter Core 1: +68.0°C (high = +100.0°C, crit = +100.0°C) $ cat /proc/acpi/thermal_zone/TZ01/temperature temperature: 27 C
This happens both in case of suspend to RAM and in suspend to disk. Any hints on things to look into would be greatly appreciated. I've since upgraded to the latest BIOS for the machine. Problem continues just as before.
Normally CPU is at 43-50C when idle. After resume from suspend/hibernate, and even when idle, temperature drifts upwards past 80C (only reboot stops the increase) What can I do to help solve this problem?
I hope it helps to point out a different but likely related problem on similar hardware is reported at: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/352583 (A900) less similar hardware, but similar problem at https://bugs.launchpad.net/ubuntu/+source/linux/+bug/352739 https://bugs.launchpad.net/fedora/+bug/234519 (the first suspend resume from power up work, but the second suspend resume fails on resume where the hd does not even wake up; requiring the power button to be held in to power down from that failed resume.) I seem to recall I had that problem as well; but don't resume anymore, so I don't see it often! What do you recommend?
Is there any information I can provide that would help move this forwards?
Perhaps the similar sounding bug https://bugzilla.kernel.org/show_bug.cgi?id=14667 has the same cause.
Apparently the fan does not come on after resume. This is a problem that others have experienced as well. There are no entries in /proc/acpi/fan neither before nor after resume. Before suspend/hibernate the fan will vary with CPU temperatures; after resume the fan stays off. Please help.
Does kernel 2.6.34.6-47 fix this problem?
How can I test that? yum update gives me 2.6.34.6-54.fc13.x86_64 which fails to boot. Which was made to appear a worse problem since the grub menu was hidden: it made it look like the system was completely unbootable. https://bugzilla.redhat.com/show_bug.cgi?id=634894 * Is there a way to install a particular kernel, or to * block a particular kernel? Your advice much welcomed here!
(In reply to comment #9) Apologies -- Comment 9 belonged on a different bug report.
In fact kernel 2.6.34.6-47 and later seem to fix the problem with the fan not coming back on * if the system resumes successfully. * However the system resumes less successfully than with previous kernels. With the new kernels, only ONE successful suspend-resume cycle is possible. It is only possible to get this one cycle if: the lid is closed without invoking the GNOME menu suspend option. Interestingly, closing the lid on the Cntl-Alt-Fn consoles does not suspend. Closing the lid on the X login window does suspend successfully once. Repeating this, or if suspend is triggered any other way, on resume the screen remains black, keyboard is unresponsive, attaching an external keyboard doesn't help as it too is unresponsive. I have opened Bug# 637650 https://bugzilla.redhat.com/show_bug.cgi?id=637650 I think I should close this bug.