On Mon, May 28, 2012 at 07:15:36PM +0200, Julien Iguchi-Cartigny wrote:
> >
> > So at least for me, it's a hardware problem,...
> >
>
> I'm not sure of this:
> - it starts around the kernel 3.0, before this everything was fine (and
> I was using this computer since 1 year and half).
> - the rather high level of Turbo mode seems strange (around 20-25% when
> idle...)
> - the presence of a similar bug in the "classical" acpi linux driver
Yup, your system is a dual-core system with the Intel graphics.
I have the quad core:
Intel(R) Core(TM) i7 CPU M 620 @ 2.67GHz
with "discrete" (i.e., Nvidia) graphics.
At least for me, the problem was also evident using a 2.6.38 kernel,
and I'm not seeing excessive Turbo mode. Unless, of course, I have an
external monitor enabled, and I do a "make -j 8" kernel build, and
then I can watch the temperature go up and up until thermal shutdown
happens.
> The best way maybe to launch such investigation in the launchpad bug
> report (which has the advantage to have a lot of people with a similar
> problem).
As an upstream developer, I find Launchpad to be a disaster, because
there are huge number of clueless users who find something that has
similar symptoms, and then say "me too!" without necessarily filling
in all of the details. As a result trying to disentangle multiple
users' bug reports, which really are for different problems, with
different kernel versions, different low-level symtoms, but perhaps
the same high-level user visible effect (i.e., "USB drives are slow")
is enough to make me pull my hair out. More than once I've just
completely given up on a Lunchpad bug report with a "there's no
intelligent life here, Scotty" reaction.
> Does anyone what are the most valuable informations to trace the problem
> for developpers ?
> - computer, cpu and graphic chipsets
> - the turbo mode percentage
> - ???
BIOS version is another really good data point (more than once these
sorts of problems were magically solved when users upgraded to the
latest firmware --- and it's another classic data point which most
Launchpad users don't bother to report).
What I'd suggest is creating a wiki page on the Thinkpad wiki that
explicitly lists the sort of information that is needed, with a very
careful description of the symptoms that you are seeing, including
things like excessive Turbo mode.
Then ask the help of people to fill in a table on the wiki page with
the relevant data points, and whether or not they are seeing the
symptoms that you are seeing (i.e., a checkbox if they see the
excessive Turbo mode).
One huge advantage of a wiki is that irrelevant information can easily
be groomed and moved to another page (if there is another bug with
similar symptoms), in a way that Launchpad simply can not do.
Launchpad gets very slow and incredibly painful to use if there are
more than 100 comments (which a bug like this is very likely to have).
Bottom line? Friends don't make friends use Launchpad. I've largely
given up scanning Launchpad for ext4 bugs, just because I've wasted
too much time. The huge amount of unqualified bug reports with
largely non-existent grooming by Canonical means that as far as I'm
concerned, it adds significant negative value to me as an upstream
developer. (And yes, if you want to help this bug get fixed, you may
need to help to keep the wiki page groomed; but at least that's
something you can do, unlike Launchpad, which doesn't support the
ability to suppress or move irrelevant/useless comments.)
- Ted
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
ibm-acpi-devel mailing list
ibm-acpi-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.sourceforge.net/lists/listinfo/ibm-acpi-devel
