Bug 508192
Summary: | sensors correctly functioning in F10 are not loadable in F11 | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | John Mellor <john.mellor> |
Component: | lm_sensors | Assignee: | Phil Knirsch <pknirsch> |
Status: | CLOSED NOTABUG | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
Severity: | medium | Docs Contact: | |
Priority: | low | ||
Version: | 11 | CC: | hdegoede, pknirsch, rvokal, tra |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | i686 | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2009-07-02 08:22:29 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Description
John Mellor
2009-06-25 23:56:21 UTC
I cannot solve the problem but as additional information I have exactly the same problem with Abit KW7 board (uses same w83627hf module). It worked fine with F10 but not anymore with F11. If I manually try to load the module I got error: bash-4.0# modprobe w83627hf FATAL: Error inserting w83627hf (/lib/modules/2.6.29.5-191.fc11.i586/kernel/drivers/hwmon/w83627hf.ko): Device or resource busy bash-4.0# Please advice how to debug this more, why this is busy? BR Tommi Hi, This become a bit of a FAQ, so I've made a blog post about this explaining what is going on (and why this is not a bug), this blog post also includes a workaround: http://hansdegoede.livejournal.com/7932.html Regards, Hans I very strongly disagree. It certainly IS a bug, and it is not appropriate to close this issue without a fix. If the kernel is now pre-empting the use of the old mechanism, then the package should be invoking the appropriate API from the kernel to do these probes instead of doing them itself. This means that the package should be redesigned to do this, and not left in a non-working state. As it sits, the package is not usable, the workaround is just-plain weird, and we're left with no sensor monitors of any kind. This is a major decrease in functionality that is highly visible. This workaround urgently needs to be placed in the Fedora 11 FAQ, as just about every user will want to implement it on every install. Hi, (In reply to comment #3) > I very strongly disagree. It certainly IS a bug, and it is not appropriate to > close this issue without a fix. > I'm sorry you don't like this, but the *old* behaviour is the buggy one, not the new behaviour. For example an asus eeepc laptops we were wrecking i2c transactions as both we and the BIOS was talking to the smbus controller at the same time. This is dangerous stuff, all it takes is one wrong i2c write to mess up the little eeproms on your ram modules which make the computer recognize the RAM's, after which your computer will no longer work and you can basically throw your RAM's away. > If the kernel is now pre-empting the use of the old mechanism, then the package > should be invoking the appropriate API from the kernel to do these probes > instead of doing them itself. This means that the package should be redesigned > to do this, and not left in a non-working state. There is nothing wrong with the package, not with the kernel, we are very *deliberately* disallowing the loading of modules which conflict with the ACPI code, if anything it is a BIOS bug, go complain to your motherboard vendor. For example Asus has this problem solved by offering an ACPI interface to get the sensors reading through ACPI (as ACPI owns the hwmon chip). This means that not only this problem does not exist with Asus boards, but hardware monitoring on them is 100% plug and play, no need to run sensors-detect, no need to tweak sensors.conf, it just works. Abit OTOH has always been notoriously bad in supporting Linux, for example their own custom Abit uGuru chips for hardware monitoring had to be reverse engineered. Trust me on this, I'm the person who wrote the Linux Abit uGuru drivers, and they aren't pretty. Regards, Hans |