Re: overcoming known kernel bug
Hi Jeroen,
Thank you for your questions. Yes, I have.
>From my reading understanding the linux kernel consists of a lineage of patches to the source code. I imagine a big ship floating because of a lot of duct tape. I expect that just applying a recent patch an old kernel would not be the intelligent thing to do.
Again, the bug is described and patched in the following thread.
https://bugzilla.kernel.org/show_bug.cgi?id=196129
Supposedly, that patch made it into the mainline kernel v. 4.13.4.
I have tried the following:
# Get stuff
kversion="4.14.2"
wget https://kernel.org/pub/linux/kernel/v4.x/linux-$kversion.tar.gz
wget https://kernel.org/pub/linux/kernel/v4.x/linux-$kversion.tar.sign
zcat linux-$kversion.tar.gz | gpg --verify linux-$kversion.tar.sign -
tar xzf linux-$kversion.tar.gz
cd linux-$kversion
Then configure and compile:
make olddefconfig
make deb-pkg
I also tried:
# From: https://github.com/nealk/thinkpad_trackpoint_linuxkernel
cp /boot/config-`uname -r` .config
yes '' | make oldconfig
make clean
make -j 3 deb-pkg
But then I run out of disk space.
I have also tried:
make nconfig
But then I am overwhelmed with all the new features.
So yes, the I failed to compile linux kernel such that it worked. I remember restarting, and then the touchscreen didn't work nor did the keyboard.
The DKH (http://kernel-handbook.alioth.debian.org/) gave me a nice schooling, but I am afraid I don't have the time to make it work in the end. Therefore, I ask for help.
I guess I should try again starting with the relevant part of the manual (http://kernel-handbook.alioth.debian.org/ch-common-tasks.html#s-common-official). I would if there is a repository of kernel config files specific to computer models and kernel versions. That would be cool and make life a little easier for those that want a custom kernel. Anyways.
In the end, I just configure my laptop to turn the screen off upon closing, when I want to keep working shortly afterwards. That's not so bad for use in public transportation. Coding there is kind of fun.
Cheers
Brian
On Wed, 2017-12-13 at 14:49 +0100, Jeroen Mathon wrote:
> Hey David,
>
>
>
> Have you tried compiling your kernel from source, or researching
> if a certain option in the kernel config could cause this?
> Interesting things to look at: Menuconfig, Kernel patches, Custom
> Drivers.
>
>
>
>
> On 12/13/2017 02:22 PM, Brian J. Oney
> wrote:
>
>
>
> > Dear David,
> >
> >
> >
> > this is different issue. My CPUs are just fine and I can use
> > the laptop without issue. Still the my laptop whines.
> > Downgrading, upgrading to a rolling release, or (gasp)
> > installing windows would be running from the problem.
> >
> >
> >
> > Ideally, a kernel shepherd would teach me to coax my sheep to
> > calm down when being rewoke. It's bleats but does not stink, and
> > all I have figured out to remedy this is to knock it out .
> > Killing this bleating lamb is not an option.
> >
> >
> >
> > Cheers,
> > Brian
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > On Tue, 2017-12-12 at 18:47 -0800, David Christensen wrote:
> >
> > > On 12/12/17 05:30, Brian Oney wrote:
> > > > I am having trouble with my 2016 lenovo thinkpad yoga 11e (3rd gen) running
> > > > the current version of debian stable (stretch). The on wake-from-suspend
> > > > the fan runs on high.
> > > >
> > > > Specifically, I have:
> > > >
> > > > ~ $ uname -a
> > > > Linux tinkbox 4.9.0-4-amd64 #1 SMP Debian 4.9.51-1 (2017-09-28) x86_64
> > > > GNU/Linux
> > > >
> > > > On wake-from-suspend:
> > > >
> > > > ~ $ sensors
> > > > thinkpad-isa-0000
> > > > Adapter: ISA adapter
> > > > fan1: 6125 RPM
> > > >
> > > > acpitz-virtual-0
> > > > Adapter: Virtual device
> > > > temp1: +65.0°C (crit = +90.0°C)
> > > >
> > > >
> > > > The acpitz-virtual-0 pegs the temperature at 65°C and won't let it go.
> > > > Therefore the fan attempts liftoff.
> > > >
> > > > I could attach the output of 'reportbug kernel', but the problem is known
> > > > and the bug is described in:
> > > >
> > > > https://bugzilla.kernel.org/show_bug.cgi?id=196129
> > > > The bug is present up until it's fix in kernel 4.13.4 or something around
> > > > that time. The solution is to install a much newer kernel (or downgrade).
> > > >
> > > > Being lazy I tried to just install the latest backported kernel
> > > > (linux-image-4.13.0-0.bpo.1-amd64). That doesn't work.
> > > >
> > > > What I find most interesting would be to compile a slimmer, faster kernel,
> > > > but I have failed (after consulting the debian kernel handbook). One thing
> > > > or the other doesn't work afterwards. Also, I run out of disk space lately
> > > > (15Gb is huge!) My idea was to use the old kernel configuration (with
> > > > 'make olddefconfig'), but there are so many new options and I honestly have
> > > > no clue how to get an overview and make an informed decision.
> > > > I would report this as a low priority kernel bug but it's (far) upstream.
> > > > It's also a known problem, which isn't necessarily debian's problem.
> > > >
> > > > I would appreciate any advice. I bought this laptop because it's tough and
> > > > has a good battery. Any laptop that misbehaves on wake-from-suspend is not
> > > > a very useful laptop (Imagine a meeting with a constantly whining laptop).
> > > > Thanks in advance!
> > >
> > > Debian 9 on certain laptops seems to have polling loop issues that
> > > manifest when the graphical login screen is displayed and when the
> > > screen saver is displayed. These are deal-breaker bugs that will burn
> > > up your CPU and suck your battery dry. Here's the bug report for my
> > > Dell Inspiron E1505/6400:
> > >
> > > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=878313
> > >
> > > The fact that it has been ignored for 2 months is not encouraging.
> > >
> > >
> > > My work-around has been to pull the battery, plug in the power adapter, run:
> > >
> > > cpufreq-set -g powersave
> > >
> > > to minimize heating/ damage when I log out/ screen lock, and run:
> > >
> > > cpufreq-set -g ondemand
> > >
> > > when I log in.
> > >
> > >
> > > I read a post somewhere that someone had found a way to muck with
> > > configuration settings and make at least some of the problems go away,
> > > but I don't have that URL.
> > >
> > >
> > > Looking at the Debian Testing kernel packages, it doesn't look like
> > > Testing includes the bug fixes you mention (?):
> > >
> > > https://packages.debian.org/testing/kernel/
> > >
> > >
> > > Ideas:
> > >
> > > 1. Go older -- e.g. Debian 8 or Debian 7.
> > >
> > > 2. Go bleeding edge -- e.g. Debian Unstable, Fedora, or Arch.
> > >
> > > 3. Run Windows and a hypervisor.
> > >
> > >
> > > David
> > >
> > >
> > >
> >
> >
>
>
>
>
>
Reply to: