On Fri, 22 Apr 2011, Alessandro Crismani wrote: > I've just bought a thinkpad T420s and I love it. The only issue I have at the > moment is with the fan control and the thermal management. In particular I > have no /proc/acpi/imb/thermal and no /proc/acpi/ibm/ecdump (even with > experimental=1). What I see is only: Yeah, Lenovo is disabling these stuff as of lately, but it could simply be lack of support for the T420 in the driver. I'd need a boot with the "debug=0xffff" parameter of thinkpad-acpi to tell you more. > I am also seeing rather high fan speeds. If the laptop is left idle the fan You could open a bug against the ACPI component in bugzilla.kernel.org, as this is a general problem (and not thinkpad-acpi-related) with the latest thinkpads. Might be related to the GPU as well. What GPU do you have in your thinkpad? It could also be on purpose, as Linux has no clue about some of the stuff inside the newer thinkpads with nvidia optimus, and doesn't do everything Lenovo's extra windows drivers do, Lenovo might have arranged to keep the fan always on. E.g. if no ATM (advanced temperature management) driver registers with the thinkpad-ec, use a more fan-happy profile. Since only Lenovo knows about the ATM interface (I only know it exists because they told me it exists, and apparently it has much more to do with the OS than anything else than telling the thinkpad that the OS is going to go nutters to not output much power), well... > spins continuously (at low speed, but it still spins, even if the laptop is > really cool). The fan starts spinning faster upon intensive tasks, as > expected, however when the task finishes (e.g. compiling kernel finishes), it > takes ages before the fan returs to spin slower. From /proc/acpi/ibm/fan the > average spinning speed is around 4000 RPM. There is some good amount of hysteresis in the fan, and if Linux is causing the thinkpad to actually output more power than Windows+Lenovo drivers can, it stands to reason it will take quite a while to cool down enough. If you have a thermal imaging camera, you can find out. Otherwise, well, try to gauge how much more power does Linux draw compared to windows in various situations, etc. > Is this expected? Should I provide more information and if yes how? Should I > write a fan control script based on the values reported by sensors instead of > relying on /proc/acpi/ibm/fan ? No. Never enable fan-control if you don't exactly know what is happening. I cannot stress this enough. Linux would have to be monitoring ALL thermal sensors for you to safely engage fan control, and 2.6.38 simply does not do that. It is not monitoring every thermal sensor in the chipset, memory modules, mini-ePCI cards, etc. I think Lenovo's windows drivers do. > PS: I've sent an email to the list 10 days ago but it is not showing up (I > wasn't subscribed). This is an unpdated one, but I think nobody got the > former. If you already received the first mail I'm sorry for the additional > noise. It is stuck in the moderating queue, I will do a run of that queue later. Emails to this list are never lost, but sometimes they get delayed a bit. Way too much spam to do away with the moderation. -- "One disk to rule them all, One disk to find them. One disk to bring them all and in the darkness grind them. In the Land of Redmond where the shadows lie." -- The Silicon Valley Tarot Henrique Holschuh ------------------------------------------------------------------------------ Fulfilling the Lean Software Promise Lean software platforms are now widely adopted and the benefits have been demonstrated beyond question. Learn why your peers are replacing JEE containers with lightweight application servers - and what you can gain from the move. http://p.sf.net/sfu/vmware-sfemails _______________________________________________ ibm-acpi-devel mailing list ibm-acpi-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.sourceforge.net/lists/listinfo/ibm-acpi-devel
