Messages in this thread | | | Date | Mon, 17 Mar 2008 21:33:30 +0100 | From | Jean Delvare <> | Subject | Re: Linux 2.6.25-rc6 |
| |
On Mon, 17 Mar 2008 20:35:24 +0100, Gabriel C wrote: > Jean Delvare wrote: > > Hi Gabriel, > > > > On Mon, 17 Mar 2008 15:34:04 +0100, Gabriel C wrote: > >> Linus Torvalds wrote: > >> > >>> Zhang, Rui (2): > >>> thermal: fix generic thermal I/F for hwmon > >> That commit broke lmsensros here ( found by bisect ). > >> ... > >> > >> Can't access procfs/sysfs file > >> Kernel interface access error > >> For 2.6 kernels, make sure you have mounted sysfs and libsensors > >> was compiled with sysfs support! > >> > >> ... > >> > >> Of course proc and sysfs is mounted and the lib has that support :) > > > > This has been reported as: > > http://bugzilla.kernel.org/show_bug.cgi?id=10259 > > https://bugzilla.redhat.com/show_bug.cgi?id=437637 > > > > The generic thermal zone device does something which is not > > fundamentally incorrect but that libsensors doesn't expect, and > > unfortunately libsensors was not made robust enough and dies instead of > > just ignoring the new unexpected device. libsensors 2.10.x is already > > fixed in lm-sensors' SVN [1] and a tentative patch is available for > > libsensors 3.0.x [2], however I am worried that kernel 2.6.25 will be > > released before any new version of lm-sensors so pretty much every user > > of lm-sensors will hit the problem if they upgrade to the new kernel. > > For this reason, I think we really should let the new generic thermal > > zone driver build as a module, and make it's hwmon support optional and > > disabled by default for 2.6.25 [3]. This will help mitigate the risk of > > massive breakage and complaints. > > > > [1] http://www.lm-sensors.org/changeset/5147 > > [2] https://bugzilla.redhat.com/attachment.cgi?id=298270 > > [3] http://lists.lm-sensors.org/pipermail/lm-sensors/2008-March/022724.html > > > > > Hmm , ok but then the situation is even worse. > > Why something *known* to break _all_ existing lmsensors setups out there > is being pushed that late in game ?
I completely agree that it shouldn't have been included that late in the release cycle, in particular when it obviously received no testing at all (it breaks all versions of lm-sensors.)
The reason why it was merged is that some people feared that application authors would add support for the new thermal zone interface (/sys/class/thermal) quickly instead of waiting for the same information to be exported through the standard hwmon interface (/sys/class/hwmon) which libsensors supports. This now appears to be a very small problem compared to breaking lm-sensors for all users out there.
> IMO that should be just reverted from 2.6.25 and pushed again in the next merge window.
This would be totally fine with me, yes.
-- Jean Delvare
| |