Created attachment 325532 [details] emerge --info I got a new laptop a while ago, and have been having the issue that the battery drains while the laptop is off. By "off", I mean a full shutdown (e.g. shutdown -h now). Over an eight hour off-period, the battery discharges around 15-20%. This seems to be a Gentoo / ACPI-specific issue, because it does not happen if I shutdown the laptop from Windows or from an Ubuntu 12.04.1 Live-CD. It also does not happen if I boot Gentoo with the "acpi=off" kernel parameter, and then perform a shutdown. I have tried various things to solve this, but have been unsuccessful: 1. Replace the battery. 2. Upgrade from kernel 3.2.12 gentoo sources to 3.4.9 vanilla sources. 3. Disable all power / wakeup related features in the laptop's BIOS. 4. Disable wake-on-lan using the command "ethtool -s eth0 wol d". 5. Disable all wake-up actions in /proc/acpi/wakeup. 6. Disable bluetooth using "echo 1 > /sys/devices/LNXSYSTM:00/device:00/TOS6208:00/rfkill/rfkill1/state". 7. Disable the wlan adapter using the FN+F8 hotkey. 8. Disable the lan-adapter completely in the kernel. Here's my kernel .config: http://pastebin.com/NRzK85bp Here's my dmesg output: http://pastebin.com/BW4fE4tG
(In reply to comment #0) > Here's my kernel .config: http://pastebin.com/NRzK85bp > Here's my dmesg output: http://pastebin.com/BW4fE4tG Please attach those files to this bug report.
Created attachment 325536 [details] Kernel Config Vanilla Sources 3.4.9
Created attachment 325538 [details] Output of dmesg
(In reply to comment #1) > Please attach those files to this bug report. Done. Please note that this problem also happened with Gentoo sources 3.2.12.
laptop manufacturer and model?
Toshiba Qosmio F755-3D350.
I finally found the cause of the issue: the command "hwclock --systohc --local" causes the battery to drain when the laptop is off. This command is executed by the /etc/init.d/hwclock script. If I set clock_hctosys="NO" in /etc/conf.d/hwclock, that command does not get executed anymore, and the battery does not drain.
Sounds like broken hardware, glad you've got it fixed. If you believe this could be fixed in the kernel (I doubt it, as you're using hwclock), then feel free to report this upstream and leave a link to the upstream bug here as a reference. Otherwise this bug will suffice as a reference point for future people experiencing the bug. Since I don't think this is a kernel bug, it might not be worth fixing and there's a sufficient workaround; I'm closing this as WONTFIX for now.